On December 10, 1983, Argentina returned to democracy after almost eight years of authoritarian rule, and since then has had free and fair elections. When Alfonsín transferred the presidential sash to Carlos Saúl Menem in 1989 it was the first time in Argentinean history that a fairly elected president from one party transferred the presidency to a fairly elected president of another party.

In April 1994 elections were held to form a Constituent Assembly. Among the many amendments to the 1853 Constitution were provisions for presidential reelection, reduction of the president's term, abolition of the electoral college system, and the adoption of a second round of voting under certain circumstances. The presidential term was reduced from six to four years, and a second round of voting will be required if no candidate receives at least 45 per cent of the vote in the first round or if the winner has 40 per cent of the vote but a margin of victory over the second-place candidate of less than 10 percent. However, the reform did not touch some of the prominent features of Argentinean electoral system - strong federalism, proportional representation (PR), see List PR, closed-list ballots, see Open, Closed and Free Lists, and a threshold of three per cent of the electoral register in each district.

Under the new constitution the president, who is chief of state and head of government, is directly elected for a four-year term by universal adult suffrage. The National Congress (Congreso de la nación) has two chambers. The Chamber of Deputies (Cámara de Diputados) has 257 members elected for a four-year term by proportional representation, with half of the seats renewed every two years. They are eligible for re-election. Prior to the reform, senators were indirectly elected for a nine-year term by the provincial legislatures. Now the members of the Senate are elected in 25 three-seat electoral districts (24 provinces and the city of Buenos Aires) for a six-year term, with one-third renewed every two years. Each of the 25 electoral districts chooses three senators directly. Two seats are awarded to the most-voted party and one to the second-largest party. Governors, Municipal Mayors, and local authorities are elected according to their provincial or municipal constitutions.

Deputies are still elected by closed lists, which means that citizens are not allowed to change the order of candidates or to cross out names on the list. Moreover, most parties use closed primaries to select and order their lists. Rank and file and party elites therefore have an important impact on legislators' behavior.

Each of the 25 electoral districts has its own electoral laws. Nonetheless, it is remarkable that eleven provinces practice the "double simultaneous vote," as in Uruguay. This law allows simultaneous intra- and inter-party competition. Political parties present several candidates who compete against one another but whose votes are added together to define which party pooled the most votes. The winner is the most-voted candidate in the most-voted party.

Two partisan and institutional features contributed to the success of Argentinean democracy from democratization in 1983 to the 1994 constitutional reform. First, the two-party system ensured that the president would have a sizable bloc of legislators in congress. Second, these legislators practiced a moderate to high level of discipline, enabling presidents to pass legislation with relative ease.

Nonetheless, four additional elements that triggered the reform of 1994 were undermining the performance of the democratic system. First, the federal government controlled the flow of resources from the central government to the provinces. Second, the capacity of the president to interfere with the judicial branch undermined the system of checks and balances. Third, the closed party lists for legislative elections produced a great deal of discomfort among citizens who claimed that legislators were more loyal to party leaders than to their constituents' problems. Finally, the abuse of presidential decrees of necessity and urgency weakened the congress' ability to check the executive.

The reform of the 1853 constitution in August 1994 was the result of an extra-parliamentary agreement known as the "Pacto de Olivos" signed between Menem and former president Raúl Alfonsín. On the one hand, Menem's major objective was reelection, and it was achieved. On the other hand, Alfonsín objectives were more diffuse and difficult to understand. In essence he wanted to give a more pronounced parliamentarian style to Argentinean politics. This is why the "chief of cabinet" was created: an official who could be removed by the congress. Neverthless, the creation of this office did not reduce the high concentration of power in the presidency.

In May 1995 President Menem secured re-election with 49.8 per cent of the vote, but a major transformation occurred in the party system: a third force, FREPASO, came in second place with 29.3 per cent of the vote, leaving the Radicals in a historically low third place with 17 per cent of the votes.

In the legislative elections of October 1997 the opposition Radicals and FREPASO built a coalition called "Alianza" in many provinces to defeat the Peronist party. As a result of these elections, not only were the Peronists defeated in the largest provinces, but they also lost control of the province of Buenos Aires, where almost 40 per cent of all Argentinean citizens live. The leader of FREPASO, Graciela Fernández Mejide, a human rights activist, became the most serious challenger for the yet-unknown Peronist party candidate for the 1999 presidential race.

The 1997 legislative elections raised an important question about Argentina's political future. The incumbent Peronist party lost almost 10 percent of its support and its majority in the lower chamber, thus it will have to strike deals with the opposition if it is to pass legislation.

Australia: The Alternative Vote System

Australia is by far the most well established and best known example of Alternative Vote (AV), see Alternative Vote, in action. The system was introduced by the Nationalist government in Australia in 1918 to replace the existing first-past-the-post (FPTP), see First Past the Post (FPTP), system after it became clear that several aligned conservative candidates all standing in the same electorate could split their vote between them under first-past-the-post, thus handing victory to the less popular but more disciplined Labor Party forces. Its introduction was thus intimately related to the need to counter the possibilities of vote splitting and to encourage and reward collaboration or coalition arrangements between parties. This ability to aggregate aligned interests, rather than divide them, has long been a (largely unrecognised) feature of Australian electoral politics, but it has not been until relatively recently that the full potential of preference distribution as an instrument for influencing policy decisions has been made clear.

There is an important difference between "full preferential" and "optional preferential" versions of AV. If the decision to mark preferences beyond the first choice is left to the voter, rather than being made compulsory, then the winning candidate must gain an absolute majority of votes in the count, but not necessarily a majority of those cast. A ballot where preferences have not or cannot be assigned to a continuing candidate are said to "exhaust". By contrast, in Australia it is a legislative requirement for all preferences to be marked to cast a valid ballot. A major consequence of this is that parties distribute "how to vote" cards to their supporters on voting day, giving them the party's preferred preference ordering for all candidates which can then be copied on to the ballot by the electors, large proportions of whom do just that.

Commentators on Australian politics historically tended to regard the alternative vote as a variation of FPTP, in most cases giving results nearly identical to that system in terms of election outcomes and the structure of party systems. Douglas Rae, for example, in his seminal work on the consequences of electoral laws, stated baldly that "the Australian system behaves in all its particulars as if it were a single-member district plurality formula". A number of other commentators have argued that preferential voting makes little difference to Australian electoral results and have not been central in determining how governments are constituted.

The common element in all these judgements is the fact that they were predominantly based on the Australian federal elections of the 1950s and 1960s where, with the notable exception of the role of the Democratic Labor Party (DLP), preference distribution had little effect on electoral outcomes. Today preferences play a much more important role in deciding the outcome of Australian elections than in previous decades. It is not possible to assume that voters' primary choice would be replicated under a FPTP system, but if it were the results of the 1961, 1969, and 1990 elections would have been reversed. The decline of what was a very stable two party system, the rise of minor parties, and the increasing influence of independent candidates have all meant that the impact of preference voting has been higher during the 1990s than at any time in the past. Table 1 sets out the proportion of seats in which a distribution of preferences has been necessary to determine the result. The second column is the percentage of seats in which the eventual winner did not lead on first preferences, and thus produced a different outcome than would have been the case under FPTP.

Table 1: Proportion of Seats Where Preferences Distributed and Outcomes Changed, 1963-1996

Election Year	Preferences distributed (%)	Outcomes changed (%)
1963	19.2	6.6
1966	25.0	4.0
1969	32.0	9.6
1972	39.2	11.2
1974	26.0	7.9
1975	18.9	5.5
1977	36.2	3.1
1980	32.0	4.8
1983	24.8	1.6
1984	29.7	8.8
1987	36.5	2.7
1990	60.1	6.1
1993	42.2	8.2
1996	39.2	4.7

As the table indicates, almost half of all seats in recent elections have been determined by the distribution of preferences, although in most cases the number of winners who "came from behind" to win a seat on preferences is small, averaging around six percent in the 1990s. Even this small amount would, however, have been enough to change government in several elections.

The most graphic example of preference votes directly affecting the choice of government occurred at the 1990 federal election, where the incumbent Australian Labor Party (ALP) was polling badly and looked to be heading for electoral defeat, and where voter support for left-of-centre parties such as the Australian Democrats and Greens reached its height. The ALP, under the influence of senior strategist Senator Graham Richardson, assiduously courted the green vote, both indirectly via interactions with the major environmental lobby groups and directly via media appeals to potential green voters, appealing directly for the second or third preferences of minor party supporters, offering policy concessions on key issues and arguing that the Labor Party was far closer to their core interests than the major alternative, the Liberal/National coalition. This strategy was markedly successful: with minor party support levels at an all time high of around 17 percent, the ALP was the beneficiary of around two-thirds of all preferences from Democrat and Green voters - a figure which probably made the difference between it winning and losing the election. This was thus a "win-win" situation for both groups: the ALP gained government with less than 40 percent of the first-preference vote, while the minor parties, which did not win lower house seats, nonetheless saw their preferred major party in government and committed to favourable policies in their areas of concern.

To see how this type of preference swapping worked in practice, one needs only examine the victory of the ALP's Neville Newell in the seat of Richmond at the 1990 federal election. Newell scored only 27 percent of the first preference vote. The coalition candidate, and then leader of the National Party, Charles Blunt won over 41 percent of first preferences, and looked set for an easy victory. However, the count saw a combination of preferences from minor parties and independents, especially the anti-nuclear campaigner Helen Caldicott, flow through to Newell and enable him to win the seat with 50.5 percent of the full preference vote.

Table 2: Example: Counting in the 1990's Federal Elections

Candidate	First Count	Second Count	Third Count	Fourth Count	Fifth Count	Sixth Count	Final Count
Gibbs (Australian Democrats)	4346	4380	4420	4504	4683	Excluded	Excluded
Newell (Australian Labor Party)	18423	18467	18484	18544	18683	20238	34664 (Elected)
Baillie (Independent)	187	Excluded	Excluded	Excluded	Excluded	Excluded	Excluded
Sims (Call to Australia Party)	1032	1053	1059	1116	Excluded	Excluded	Excluded
Paterson (Independent)	445	480	530	Excluded	Excluded	Excluded	Excluded
Leggett (Independent)	279	294	Excluded	Excluded	Excluded	Excluded	Excluded
Blunt (National Party)	28257	28274	28303	28416	28978	29778	33980
Caldicott (Independent)	16072	16091	16237	16438	16658	18903	Excluded

Newell won the seat because he was able to secure over 77% of Caldicott's preferences when she was excluded at the seventh count. Caldicott herself had received the majority of preferences from the other independent candidates. The ALP in Richmond, as in other seats, was thus the beneficiary of a strategy aimed at maximizing not just its own vote but at maximizing the preferences it received from others: the "second preference" strategy. As support for the Australian Democrats and green parties reached its height in 1990, so the ALP's assiduous campaigning for second preferences saw it receive around two-thirds of the preferences from these parties, which proved decisive for their electoral victory.

The success of the ALP's strategy in 1990 was notable not least because historically the process of preference transfers has tended to benefit the non-Labor parties rather than the ALP. AV has had two main positive effects on the non-Labor parties: it facilitated the coalition arrangement between the Liberal and Country (now National) parties by enabling them both to stand candidates in some seats without the danger of vote splitting, and it enabled the preferences of one small party, the DLP, to flow predominantly against the ALP and hugely assist the coalition maintain government in the 1960s. In the 1970s, the ALP advocated a return to first-past-the-post (FPTP). But when the ALP regained office in 1983, its policy was to retain AV, but make the expression of preferences optional rather than compulsory.

Optional Preferential AV

Optional preferential AV is identical to full preferential AV except that voters are not required to express a preference for every candidate; if they wish, they can express a preference for only one. In the words of former Prime Minister Gough Whitlam, optional preferential voting is "perhaps the only electoral procedure in the world which allows electors to express their indifference to candidates. "A national survey in 1979 showed that the majority of Australian electors favoured the optional version, with 72 percent for optional and only 26 percent favouring compulsory preference marking. One clear advantage of the optional version is that the problems of spoilt ballots due to numbering errors associated with the full preferential system are largely removed. For this reason, optional preference marking is probably the only form of AV suited to conditions of low literacy or numeracy.

Optional preferential AV is currently used for state elections in New South Wales, where it was introduced by the Wran Labor government in 1981, and in Queensland, where it was introduced in 1992 on the recommendation of the Electoral and Administrative Review Commission, who considered that full preferential voting forced voters to express preferences for candidates about whom they may know little or nothing. The rate of "plumping" for one candidate only, without marking subsequent preferences, has increased over time in both these cases. The decision to express preferences also appears to be closely related to the recommendations made by parties on their "how-to-vote" cards. In a survey conducted at two by-elections in 1992, fully 75 percent of electors followed party voting directions, resulting in plumping rates of 43 percent in one district (Gordon) and 63 percent in another (Kuring-gai). In the Kuring-gai case, less than 33 percent of electors filled in all squares on the ballot paper. In Queensland, plumping rates stood at 23 percent at the first OPV election in 1992, but were significantly higher in those cases where how-to-vote material from one of the major parties did not suggest preferences. There is also a clear partisan component to plumping rates, which reflects the long-standing coalition arrangements between the Liberal and National parties: in both NSW and Queensland. Labor voters are considerably more likely to "plump" than supporters of the coalition parties.

The Effects of AV

In Australia, interest in preferential voting tends to increase with its perceived partisan effect. The influence of preferences on electoral outcomes has clearly increased in recent decades and played a crucial role in the 1990 Labor victory in particular. The collapse of the Democrat vote in 1993 and the Coalition landslide at the 1996 federal election has meant that the effects of preference distribution have received less attention since then, although it has facilitated the election of increasing numbers of independent candidates (two in 1993, five in 1996), most of whom win their seats by overtaking major party candidates on preferences.

Analyses of the effects of AV in Australia have tended to concentrate almost exclusively on its partisan impacts. Some commentators have seen the system as an instrument for maintaining the dominance of the two major parties, the ALP and the Liberal/National Coalition, and for restricting the role of minor parties in the lower house to one of influencing the policies of the major parties rather than gaining election themselves. Others claim that it can enhance the power and position of minor parties, especially if they have the potential to hold the balance of power between two major parties.

There is widespread agreement that AV has facilitated coalition arrangements such as that between the Liberal and National parties, and that it works to the advantage of centre candidates and parties, encouraging moderate policy positions and a search for the "middle ground". The sometimes fiery and aggressive rhetoric of Australian politics has often distracted observers from recognising just how much co-operative behaviour there is between parties - via preference swapping deals, for example - and how close the major parties are on most substantive policy issues. There is little doubt that the AV electoral system provides a significant institutional encouragement for these centrist tendencies.

Notes:

[1] Source: Hughes, C.A. and Graham, B.D. (1968), A Handbook of Australian Government and Politics, 1890-1964, Australian National University Press, Canberra; Hughes, C.A. (1977), A Handbook of Australian Government and Politics, 1965-1974, Australian National University Press, Canberra; Hughes, C.A. (1986), A Handbook of Australian Government and Politics, 1975-1984, Australian National University Press, Sydney; Hughes, C.A. (1997), "Individual Electoral Districts" in C. Bean, S. Bennett, M. Simms and J. Warhurst (eds), The Politics of Retribution: the 1996 Australian Federal Election, Allen & Unwin, Sydney, pp. 166-167.

[2] Source: Australian Electoral Commission, 1990 Election Statistics, AEC, Canberra.

Bolivia: Electoral Reform in Latin America

Bolivia's democratic experience has been characterized by the search for ways to solve the basic problem of Latin American presidential regimes, which have regularly slipped into stalemates between executives and legislatures led by minority governments. Most presidential systems in Latin America pose the fundamental problem that they are embedded in multi-party systems with proportional representation; this has been defined as the "difficult equation of presidentialism", and has been a permanent source of political conflicts which has adversely affected the chances of democratic consolidation.

In Bolivia the problem has been partly solved through a basic institutional shift from "presidentialism" with minority governments to a "parliamentarized presidentialism" based on majority governments. This distinctive system of government is a "mestizo child", with both parliamentary and presidentialist features. It is presidentialist because the president serves for a fixed term and, even though chosen by Congress, does not depend on its continuing confidence. But it is "parliamentarised" because the president is chosen by the legislature on the basis of post-electoral bargaining, so ensuring majority legislative support and the compatibility of executive and legislative powers. The mainspring of the system is a dynamic common in parliamentary regimes: the politics of coalition.

Like parties everywhere, Bolivian parties strive to maximize their respective vote shares, but they do not expect popular balloting to be the last stage of arbitration. Rather, they focus on post-electoral bargaining, and it is this that will determine who actually ends up in the congressional majority and with the executive power. The dominant pattern has been that of coordinated congressional and government coalitions, which has enhanced both the stability of the executive authority and the compatibility of executive and legislative powers.

Since the resumption of "free and fair" elections in 1979, the Bolivian party system, which evolved from a highly fragmented one to a moderate multi-party system of six effective parties, has proved unable to produce a single predominant party, or even alternating majorities. Thus, Article 90 of the Constitution, the guiding principle for the electoral system, has defined the normal method for choosing the president. It makes no explicit provision for political pacts, but it is its requirement that presidents be chosen by Congress when no single candidate wins a majority of the popular vote that has created broad scope for bargaining and coalition-building among political parties.

One key dimension of Bolivian "parliamentarised presidentialism" is the List PR electoral system. In fact, throughout the 1980s and early 1990s the electoral system helped reinforce the patterns of inter-party competition and coalition building, but the system also had many shortcomings and was prone to fraud and manipulation. One of the crucial issues of democratic stability and legitimacy has been the establishment of coherent rules of the game. The Bolivian electoral reforms in 1986, 1991, and 1994 were characterized by short-term calculations and contingent reactions to political pressures, and not by research or deliberate political engineering. Moreover, party leaderships lacked experience and were unable to develop a coherent reform strategy. The result was that the elections in 1985, 1989, and 1993 were all held under different PR formulas. The D'Hondt formula, introduced in 1956, was replaced in 1986 by a so-called double quotient of participation and allocation of seats, which hindered the access of small parties to Congress. In 1989 a further change established the Sainte-Lagu formula for the presidential and parliamentary elections in 1993, which encouraged, in turn, the representation of very small parties.

Nevertheless, the first wave of weighty changes had paradoxically less to do with the change of the prevailing PR system than with the establishment of an autonomous Electoral Court, the adoption of on-site vote validation of ballots at polling places, and the abolition of mechanisms that made it possible for regional electoral courts to distort results. However, the constitutional reform of August 1994 introduced a second wave of changes, and brought about the most major shift in the PR system so far by introducing, with some modifications, the Mixed Member Proportional (MMP) electoral system of Germany and New Zealand. At first this revision led to the "contradictory" adoption of parallel First Past the Post (FPTP) and PR systems - basically, a mixed PR system in terms of voting criteria but not in terms of outcomes.

Thus in August 1996, Congress had to pass a new law concerning the application of Article 60 of the Constitution to remove some obvious defects. It re-established the D'Hondt formula of PR and created a three-percent threshold for seats in the Chamber of Deputies. Henceforth, 68 deputies out of a constitutionally-fixed number of 130 will be chosen by FPTP voting in single-member districts, while the remainder will be chosen by party list voting according to proportional representation, in nine regional multi-member districts. Unlike Germany and Venezuela, there is no provision for additional seats. Seats are allocated directly to candidates winning in single-member districts, even if a party wins in only one district and obtains no PR seats. As in Germany, the overall distribution of seats, however, will be decided by applying the PR formula in a compensatory fashion, with a three-percent threshold for representation at the national level. If a party wins 10 seats through the overall List PR voting, and five seats in single-member districts, it is ultimately entitled to ten parliamentary seats.

The most striking phenomenon in the Bolivian experience of electoral reform has been the use of democratic procedures and mechanisms. Reforms were discussed in multi-party commissions and reaching multi-party consensus was a sine qua non condition for congressional approval. No referendum was called because the Bolivian Constitution does not allow this mechanism of legitimization. From 1989 through 1992, inter-party debate unfolded around two key proposals, which were, in turn, rejected. The Acción Democrática Nacionalista and the Movimiento de Izquierda Revolucionaria advocated plurality for presidential elections, so that the Congress would only have confirmed the candidate winning the plurality of votes; meanwhile, the Movimiento Nacionalista Revolucionario (MNR) proposed the French-style Two-Round (majority-runoff) System (TRS). Both proposals started from the premise that the congressional election of presidents through party bargaining did not respect the people's will, and decisions were taken behind closed doors; people voted, but did not choose the president.

A consensus was finally reached based on the MNR's proposal to adopt an MMP system for the legislature and, furthermore, to reduce the number of presidential candidates able to obtain a plurality of votes at the parliamentary election from three to two, and to establish a five-year mandate for the president, the vice-president, and members of parliament. The real shift to MMP-style PR stemmed from discontent with vote manipulation in the 1989 general election, but the specific causes of the reforms were three-fold: the concern about a process of de-legitimization of party representation because closed party lists weakened the links between MPs and voters; the disillusionment of citizens with a lack of political responsiveness and accountability of governing parties; and finally a desire to reduce the growing alienation between parties and society by fostering constituency representation.

In the presidential and parliamentary elections of June 1997, these electoral reforms did not have the effects expected, as the party system became more fragmented and polarized than the one elected in 1993. For example, in 1993 the largest party won 35.6 percent of the vote; in 1997, the largest party - a different one - won only 22.3 percent. Only seven parties won seats in 1997, compared to nine in 1993, but the delegations were much more equal in size, making for a significantly more fragmented congress. There were three reasons for this unexpected outcome. First, the National Revolutionary Movement (MNR) of incumbent president Gonzalo Sanchez de Lozada lost nearly half of its share of the vote, depriving it of its temporary dominant position vis-a-vis its competitors. Second, in 1993 the MNR's two principal rivals, AND and MIR, were joined in an alliance called the Patriotic Accord; before 1997 this pact broke apart, and ADN and MIR ran separate presidential candidates and presented separate congressional lists. It is tempting to argue that there would have been fewer parties if these two events had not transpired; however, the MMP electoral system actually appears to have worsened the fragmentation. Due to the unusually high degree of regional concentration of party support, more parties (seven) won seats in the new single-member districts than in the multimember districts (five parties). Overall, the new parties were more personalist than before, but it is difficult to attribute this outcome to the mixed electoral system, as many of the personalist deputies were elected by PR.

Brazil: Candidate-Centred PR in a Presidential System

In 2002 Brazilians went to the polls to choose a new president, the members of the bicameral national legislature, governors for the component parts of the federation (26 states plus the Federal District of Brasília), and members of the unicameral state legislative assemblies. This was the fourth direct election since the end of the military regime in 1985 of the president and all other major legislative and executive posts.

Presidential elections in Brazil take place under a two-round majority run-off system, with candidates competing for votes throughout the country’s 8,511,965 sq km area. Following a constitutional amendment approved in June 1997, presidents are now allowed to run for re-election once. Fernando Henrique Cardoso, the incumbent at the time the amendment was approved, won re-election in 1998 in the first round with 53.1 per cent of the vote. However, Luiz Inácio Lula da Silva polled 46 per cent in the first round in 2002 and was elected in the run-off round.

The rules governing legislative elections have remained essentially unchanged since they were first established in 1946. The Senate is the chamber where the regions of Brazil are represented: each of the 27 component parts of the federation is represented by three senators who are elected by plurality for an eight-year term. Membership is renewed every four years by one-third and two-thirds, in alternation: when two senators are to be elected, voters have two votes under a Block Vote (BV) system.

The Chamber of Deputies has 513 members who compete in 27 multi-member electoral districts, corresponding to the 26 states and Brasília. Their magnitude is determined by population, subject to the restriction that no state can have fewer than eight or more than 70 representatives. Elections take place under a system of open-list PR. Each voter has one vote to cast, which can be given to a political party or to an individual. Votes given to candidates from each party are pooled and added to the votes received by that party to give a total party vote, which is used to determine the number of seats to be allocated to each party. The candidates with the most votes on each party list win the seats allocated to that party. Seat allocation has been made under the D’Hondt Formula since 1950. Parties that do not gain a full quota in a district are, however, excluded from gaining a seat. Until 1998 the calculation of the quota was based on the total number of valid and blank votes, making the threshold for representation higher.

Deliberate Malapportionment

The rules for the Chamber of Deputies elections are probably the most controversial element of the Brazilian electoral system. The floor and the ceiling on the size of electoral districts mean that representation in the Chamber in terms of population is uneven across the states. This seriously violates the principle of ‘one person, one vote, one value’ (OPOVOV), as the number of votes necessary to elect one representative in São Paulo, which has over 25 million voters and 70 seats, is ten times higher than it is in Amapá, which has about 290,000 voters and eight seats. The resulting malapportionment benefits the less populous states, which tend to be poorer and more reliant on agriculture, and is disadvantageous to the larger states, which are richer and more industrialized. For this reason it has been blamed as one of the main mechanisms for reinforcing traditionalism in politics and thereby weakening political parties.

However, this needs to be qualified. The only significant loser from malapportionment is the state of São Paulo, where the number of representatives would increase by about 40 if the size of the electoral districts reflected population size strictly. Some other states are marginally under-represented, the second-biggest loss occurring in Minas Gerais (about four representatives). The losses due to malapportionment are therefore concentrated. They also reflect the goals of the makers of the 1946 constitution, who were concerned with finding a formula that would prevent São Paulo (and to a lesser extent Minas Gerais) from dominating the federation as they had done during the period known as the First Republic (1899–1930).

To the extent that malapportionment favours relatively poor states politically, it may help to promote a regional redistribution of wealth that is of no small consequence in a country with such high levels of regional inequality as Brazil.

In addition, the frequent assumption that over-represented states are capable of systematically blocking legislation of national scope remains to be proved. It is not necessarily the case that the pattern of politics that characterizes the over-represented states is any different from the one in the under-represented ones. Clientelistic practices exist in all states, and elections are mass phenomena that generate a high degree of competition. If clientelism characterizes Brazilian politics, malapportionment of the Chamber of Deputies is unlikely to be a significant cause.

Competition Between Parties—and Within Parties

One of the main features of the system of open-list PR for the Chamber of Deputies is that it induces both inter- and intra-party competition. These elections are quite competitive. For example, in 2002 a total of 4,901 candidates stood for the 513 seats in the Chamber. In only nine of the 27 districts were there fewer than 100 candidates; the lowest number was 66 for eight seats in Tocantins. There were 793 candidates for 70 seats from São Paulo, 602 for 46 seats from Rio, and 554 for 53 seats from Minas Gerais. Parties compete with each other. Candidates, seeking to be elected for the seats which their parties gain, compete among themselves for the votes their parties obtain. This is said to lead to personalism, which is considered to be at the root of the weakness of Brazil’s political parties, to clientelistic ties between voters and their representatives, and to a national legislature that is primarily concerned with local rather than national, and clientelistic rather than programmatic, issues.

Again, this view needs to be qualified. First, the view that it is personalism that mainly drives voters’ decisions in elections to the legislature in Brazil is far from well established. Although the proportion of preference votes (when the voter chooses a specific candidate, not simply the party) is far larger than the proportion of party votes, these figures say very little about how voters actually decide. If voters give greater relative weight to the individual than to the party, many voters who vote for a specific candidate would presumably also vote for that candidate even if he or she were to change parties. While no studies have tried to address this issue directly, scattered evidence indicates that representatives who switch parties in the middle of the legislative term are less likely to be re-elected, which suggests that they are not able to carry with them the votes that got them elected in the first place.

Voters and Their Representatives

Even less is known about the ties between voters and their representatives. A great deal of effort has been spent trying to uncover the pattern of clientelism and localized favours that must have served as the basis for a successful electoral campaign and legislative career. Successful candidates, it is said, are those who bring ‘pork’ to their ‘constituency’. In Brazil’s multi-member district system, however, the individual member is one of at least eight representing the district, which makes it difficult to establish the link between a particular member and a new spending project. Even though some candidates may and do try to carve de facto geographic constituencies for themselves, this is not the only, and may not even be the most effective, way of getting into the Chamber of Deputies. One study of the geographical distribution of the votes of successful candidates demonstrates that in 1994 and 1998 only about 17 per cent of representatives adopted such a strategy, that is, were able to obtain the largest share of votes in a cluster of geographically concentrated localities. The others adopted different strategies, such as sharing with competitors a relatively defined geographic area, dominating localities that were distant from each other, or obtaining relatively small shares of their total vote in geographically dispersed areas. Given the level of competition of elections and the lack of legally protected constituencies, it is unlikely that a representative will feel safe about his or her ‘bailiwick’. Indeed, rates of re-election are not very high: estimates put it at around 60 per cent of those who seek re-election. Thus, clientelism does not characterize, at least not exclusively, the ties between representatives and voters.

Does the Electoral System Contribute to Party Fragmentation?

There is much we still need to know about the way in which the system of open-list PR with large electoral districts, such as the one that exists in Brazil, operates. We do know, however, that elections are extremely competitive, that the advantage of incumbency is relatively weak, and that deputies’ relations with their electoral districts differ, so that there is no dominant strategy for a successful candidacy.

The extent in which the electoral system induces clientelism and individualism inside the Chamber of Deputies is at least questionable. While it is beyond the scope of this overview to discuss the mechanisms which the president and the party leaders may use to shape the behaviour of individual deputies, deputies face other pressures in addition to the demands of localized and particularistic constituencies. These pressures are a counterbalance to increased party fragmentation.

Party fragmentation in the Brazilian legislature has been held responsible for a number of the malaises the country has suffered from in the past 15 years. The high degree of fragmentation of the party system is usually attributed to a combination of factors, which include the electoral system and its individualistic tendencies, the characteristics of presidential systems, and the strong federalism adopted by the 1988 constitution.

The degree of fragmentation in the Chamber of Deputies has, however, remained constant, at around eight effective parties, since the 1990 election. Some aspects of the electoral law tend to favour the larger parties and work against fragmentation. Examples include the adding of blank votes to the base on which the electoral quota is calculated (which makes the quota larger and hence more difficult to achieve), and the exclusion of all parties that do not obtain one quota in a district from winning a remainder seat.

The links between presidentialism and party systems are not yet well enough understood. This leaves federalism as a possible cause of fragmentation of the party system. Some of the national parties in Brazil are de facto coalitions of regional parties. Smaller parties emerge out of these coalitions for purely local reasons, thus leading to a multiplication of parties at the national level. Whether this is the real or the only reason why new parties emerge, it remains unclear whether federalism is a cause of fragmentation or simply a reflection of the variety of regional interests that a country as large as Brazil must accommodate in order to operate democratically.

Canada: The Canadian Electoral System

When three of the remaining British colonies in North American federated in 1867 - the same year Britain extended its suffrage to 10 percent of the electorate - the new Dominion of Canada naturally adopted British institutions of electoral democracy. Canada's founding fathers, in contrast with their Australian counterparts two generations later, failed to ask if the British First Past the Post (FPTP) system was suited to a federal country dispersed over far-flung regions. Though some local and provincial experimentation with different systems of election took place after the Western provinces entered confederation earlier this century, it proved short-lived. Today, not only are the 308 Members of Parliament elected through FPTP, but so are all members of the ten provincial legislatures and three territories. Indeed, over the years, the federal electoral system moved even more closely to the pure FPTP plurality model as the few two-member districts that existed were gradually eliminated.

That FPTP is appropriate for Canada has largely been taken for granted in part because Canadians' familiarity with electoral experiences outside its borders generally extends only to the US and UK. Yet, this does not fully explain how a country so much concerned with constitutional reform has not proven open to altering its electoral institutions - especially, as we shall see, given the anomalies they have produced. This is not to say that reform to a more proportional system has never been proposed; only that it has not made it to the political agenda. The Task Force on Canadian Unity (Pepin-Robarts Commission) in its 1979 Report included a recommendation for just over 20 percent of the seats in the House of Commons to be accorded to the parties proportional to their support and from those provinces in which there was underrepresentation. A slightly different proposal was submitted by the left-leaning New Democratic Party, the party most underrepresented under FPTP. Yet when the Trudeau government rejected the Pepin-Robarts report, electoral reform of the House of Commons was also shelved.

The fact that the issue was off the political agenda became clear ten years later when Pierre Lortie, Chairman of the Royal Commission on Electoral Reform and Party Financing set up by the Mulroney government in 1990, made it clear that changing the electoral system as such was outside the Commission's mandate. Discussion of electoral reform of federal legislative institutions concentrated on a proposal supported by the Western provinces to turn the appointed upper chamber, the Senate, into an elected one. But when Senate reform died with the rejection of a Constitutional amendment proposal in a 1992 referendum, this possible avenue to electoral regimes other than FPTP was closed.

Ironically, the distorting effects of the FPTP electoral system on representation in the House of Commons - combined with Canadians' tendency to identify politically along regional lines - have probably never been greater than in the two federal elections that took place in the 1990s. In 1993, the voters repudiated the ruling Progressive Conservatives, but the electoral system almost decimated Canada's oldest party. Rather than electing the 46 members of 295 that a proportional system would have given them, the Tories managed to elect only two. In contrast, the two regionally-based parties, the Bloc Québécois and Reform, with 13.5 and 19 percent of the popular vote respectively, elected 54 and 52 MPs.

In 1997, of the 301 seats in Parliament, the Liberals won 155, Reform 60, the Bloc Québécois 44, the NDP 21 and the Tories 20. Had the seats been distributed according to the parties' popular support, the Conservatives would have placed third with 58 seats, just behind Reform's 59, with the NDP up and the Bloc Québecois down to 33 each, leaving the Liberals with 118. Two thirds of the Liberals' seats came from Ontario, while Reform dominated the Western provinces, and the Bloc Québécois Québec - "quartering Canada" - as The Economist put it, producing what Canadian pundits called a "Rainbow Parliament." Had the seat been distributed according to the popular support for the parties, Liberals, Conservatives, and NDPers would have won seats in all provinces or regions; Reformers in all but Québec. And this, of course, is to leave out the fact that under PR the parties would have had an incentive to expend their efforts and resources beyond the regions where they do well: the Conservatives would have put far more effort into the West; the NDP and Reform would have worked much harder for support in Québec. Indeed, there is good reason to assume that the low turnout of just over two-thirds of registered voters is linked to the fact that in most ridings only one or two of the parties were real contenders, with supporters of the others effectively disenfranchised.

Electoral reform toward a more proportional system was proposed by a number of columnists and editorialists in the wake of the two elections, and raised by the leaders of the Progressive Conservative Party, but only wistfully. And in November 1997, a private member's bill was submitted by a leading member of the NDP proposing Parliament endorse PR and appoint an all-party committee to conduct public consultation on the question and report back with a concrete proposal which would then be put to Canadians for their approval a national referendum. Yet, like other private members' bills, this one will die on the order paper. By and large, politicians view electoral reform as a non-starter in which they are unwilling to invest precious political capital.

While this is understandable it is also regrettable. While the FPTP system has produced some majority governments, the tendency of the system to polarize rather than promote compromise has not necessarily served Canada well. As a thought experiment, one can imagine the outcome if the one serious recent effort to bring electoral reform to a provincial political agenda had succeeded. This was in Québec in the early 1980s when an investigatory commission advocated adoption of a regional-list system of PR, a recommendation endorsed by the Québec cabinet but one that due to lack of support from the opposition, and even in the governing party caucus - was never presented to the legislature. Had it been adopted, the balance of power today would be held by parties representing the twenty-five percent of Québeckers who want change but prefer a compromise short of the sovereignty favoured by the Parti Québécois.

The only electoral reform efforts that made it to the political agenda were provisions adopted in certain Western provinces allowing for the recall of legislators. As far as electoral-system reform is concerned, the only real prospect might be for Canada to once again follow Britain's example. If Britain proves prepared seriously to consider changing the electoral system it bestowed upon Canada, Canadians might follow suit.

British Columbia: Empowered Citizen Participation

The government of the Canadian province of British Columbia, with the full endorsement of the province’s Legislative Assembly, has initiated a historic, unique and precedent-setting process on electoral reform by establishing the Citizens’ Assembly on Electoral Reform. This is the first time that a government has given a randomly selected group of citizens the opportunity and responsibility to independently review the electoral system and have its recommendation submitted to the public for approval at a referendum.

The 1996 election for the British Columbia provincial legislature was conducted under an FPTP system. It resulted in the New Democratic Party (NDP), with 39 per cent of the popular vote, winning 39 seats in the Legislative Assembly—more than the 33 seats gained by the Liberal Party, which had won 42 per cent of the popular vote. The NDP, with less popular support than the Liberal Party, thus formed the government for the next five years. This result motivated the Liberal Party to make electoral reform a priority in its political campaign for the next election. At the 2001 election the Liberal Party promised to implement electoral reform through a Citizens’ Assembly: following an election victory which gave it 97 per cent of the seats in the legislature with 58 per cent of the popular vote, it clearly had the mandate to pursue these objectives.

The typical approach used in Canada for the development of public policy issues where the government is seeking public review is to establish a commission or board of public inquiry, usually led by judges, experts or political leaders. After inviting submissions from the public, and following a period of wider consultation, the government makes a decision on the actions that will follow, taking into account the report produced by the commission.

The blueprint of the Citizens’ Assembly and the framing of its terms of reference were prepared by Gordon Gibson, an author on democracy and former political party leader active in business and public affairs, and the new government in consultation with electoral reform experts. There were two unique and precedent-setting features for British Columbia: the people appointed would not be experts or specialists in the field of electoral reform, but would instead be randomly selected citizens from across the province; and, if a change were recommended, the question would be put directly to the citizens of the province at a referendum and would not be filtered through the government.

The Citizens’ Assembly that resulted was a non-partisan and independent group of 160 men and women of all ages from across the province of British Columbia, chosen by random selection from the electoral register. The selection phase was designed to give a balanced list of men and women, reflective of the age distribution of the population of British Columbia as reported in the 2001 census, including two members from the aboriginal community, and representing the whole of the province. This was followed by an intense learning phase for the Assembly during which various electoral system experts produced learning materials (all also available to the general public) and held sessions with the members to inform them of the different systems available and discuss their advantages and disadvantages.

At the conclusion of the learning phase a report, Preliminary Statement to the People of British Columbia, was sent to various groups in society, including members of the Legislative Assembly, libraries, municipal district offices, schools and universities, to inform the public of the preliminary conclusions of the Citizens’ Assembly. This report was followed by a phase of public hearings, during which about 3,000 people attended some 50 hearings held in all areas of the province. During the subsequent deliberation phase, plenary sessions and discussion groups were held at which the Assembly narrowed down the choice of electoral systems to two and, as a group, sketched out the details of each system. The first day of that phase featured a repeat of some of the best presentations heard during the public hearings—presentations that advocated a variety of electoral systems and features. The objectives of all these phases were to identify the elements essential to a British Columbian electoral system, review thoroughly all electoral system options in the light of these elements and, most importantly, to increase public awareness, inclusion and participation. The three essential elements arrived at in the end were voter choice, local representation and proportionality. Finally, in late October 2004, the Assembly presented its recommendation, in which it supported (by 146 in favour to seven against) changing the FPTP system to STV. The completion of the Citizens’ Assembly process then required the publication of the formal final report and the submission of the recommendation to referendum.

This participatory model attracted significant interest from groups across Canada. The concept was recommended to other governments within Canada as a good way of involving citizens in issues that should be the domain of citizens, and a similar process to the one in British Columbia was initiated by the Ontario government.

Other elections in Canada have also contributed to the growing support for a review of electoral processes. Federal majority governments have often been elected with significantly less than 50 per cent of the popular vote. As a result, a number of initiatives for a change of the electoral system at federal level, including Fair Vote Canada (FVC), have emerged, as have many individual lobbyists and advocates.

There is reason to think that the experience with the British Columbia Citizens’ Assembly will have significant impact on the future of the debate on electoral system change, and on the process of review and change in particular, on a federal level in Canada. Following pressure from both the NDP and the Conservative Party, the following amendment to the Speech from the Throne was unanimously accepted in October 2004: ‘an Order of Reference to the Standing Committee on Procedure and House Affairs instructing the committee to recommend a process that engages citizens and parliamentarians in an examination of our electoral system with a review of all options’.

The future impact of the British Columbia Citizens’ Assembly on the process of review and change of electoral systems on an international scale remains to be seen, but it is safe to say that its establishment and work have raised interest in and added to the empirical knowledge of participatory processes around the world.

Switzerland

The Swiss parliament has two chambers, the National Council and the Council of States. In the National Council, the cantons are represented according to population. In the Council of States, each canton has two representatives, but there are also a few half-cantons with one representative each. For the National Council, there are uniform electoral rules for the country at large; for the Council of States, it is up to each canton to determine the electoral rules as long as they are democratic.

When modern Switzerland was founded in 1848, the electoral rules for the National Council were winner-take-all in single member districts such as FPTP, see First Past the Post (FPTP). After World War I, the rules were changed to party list proportionality, see List PR. Currently, the National Council has 200 members that are elected in 26 electoral districts, corresponding to the 26 Swiss cantons and half-cantons. The largest canton, Zurich, elects 35 representatives, and the smallest cantons, only one. The parties submit candidate lists in each canton containing the names of their candidates for that canton's seats. The results are counted separately for each canton.

Having 26 electoral districts instead of a single national district works against the smaller parties. If Switzerland were treated as a single electoral district, only one-half of one percent of the vote would be needed to win one of the 200 National Council seats. With elections taking place in 26 separate districts, however, a higher percentage of votes is needed to win. In Zurich, a party must win about three percent of the vote to win one of the canton's 35 seats. In the small cantons with only one seat, the party with the most votes wins the seat. Thus, if the number of seats per district is reduced to one, the proportionality system becomes a system of winner-take-all (FPTP).

In contrast to countries like Germany, Switzerland has no minimal threshold of votes that a party must reach to receive any representation at all. Thus the principle of proportionality is applied in its pure form.

The candidates on the party lists are ranked by the voters and not by the parties. The latter merely submit a list of names without rank, usually in alphabetical order. The number of names cannot be more than the number of seats to be filled from each respective canton. In ranking the individual candidates the voters have three options:

Leave the candidate’s name, as it appears on the list,
Put the candidate on the list a second time, or
Drop the candidate from the list.

The only condition is that the overall number of names is not greater than the number of seats to be elected from the canton. A voter also can decide to make no changes at all on the list. In this case, no preference is given to any of the candidates, but the ballot counts for the number of seats attributed to the party.

Voters may further complicate their list by writing in candidates from other parties (panachage). Thus, a Socialist voter may put a Free Democratic candidate on the list, either once or twice. With this write-in possibility, computation of the results becomes very complicated. In the above example, the Free Democratic write-in candidate counts for the Free Democratic party and detracts from the Socialist party strength; the voter has split their vote between the two parties. Voters can go even further and write in candidates from as many parties as they wish, but, again, the total number of names is not allowed to exceed the number of seats in the canton.

The computation of the results proceeds in the following way: for each canton, the number of seats each party receives is determined on the basis of the total votes for candidates of this party. Second, candidates win these seats in order of their ranking. This ranking is based on the number of times a candidate's name appears on all the lists, including write-ins on other parties' lists.

The freedom of choice that the Swiss system permits the voter weakens the party's control over its candidates, and thus party discipline is low. Although a Swiss party still controls whether or not a candidate gets listed, it cannot determine a candidate's chances of election through rank on the list. Once candidates are listed, they are on their own and must try to get a maximum number of voters to write them in twice, and a minimum to cross them out. Although this system seems to give great power to the electorate, it also increases the influence of interest groups. These groups inform their members about the candidates who favour their interests and for whom two votes should be cast, as well as about candidates who should be crossed out because they do not favour the group's interests. A teachers' group, for example, will inform its members which candidates are sympathetic to teachers' needs and which are not. Letters are sent out by a large number of groups ranging from business groups to trade unions. Candidates depend on political parties only for getting listed on the ballot; to be elected, they must obtain the support of a large number of different interest groups.

The Swiss still vote for party lists, but their electoral system allows them to express preferences for and against particular candidates. The election also takes place in relatively small districts, where voters feel more at home than in a single national district. These factors together personalize the relations between voters and candidates. With this electoral system the Swiss currently have 14 parties in the National Council.

Besides taking part in elections, Swiss voters have also a great say through the referendum, see Referendum. Indeed, of all the national referenda held in Western democracies since World War II, more than two-thirds were held in Switzerland. Voters have the right to call for a popular referendum on every bill decided by parliament. The only requirement is that 50,000 signatures be obtained, which is relatively easy in a country of 7 million inhabitants. The voters also have the final say on constitutional amendments. All constitutional amendments decided by parliament must be submitted to the voters. A minimum of 100,000 voters can also submit a constitutional amendment of their own, which will first be debated by parliament but finally decided in a popular referendum. This instrument of the popular constitutional initiative is widely used and can be applied to whatever question the people wish to decide. When the referendum was introduced in the 19th century, it was expected that its effect would be innovative. The founders of modern Switzerland anticipated that the voters would be open to change, but the opposite was true and the referendum has often had a delaying effect. The best example is the introduction of female suffrage only in 1971. Parliament was prepared much earlier than male voters to grant women the right to vote. This example is typical in the sense that it shows how it often takes a long time to convince the Swiss voters to accept a new idea.

Swiss Direct Democracy

A detailed overview of the numerous direct democracy mechanisms used in Switzerland

Direct democracy was introduced at federal level in Switzerland in 1848, although in some Swiss cantons forms of direct democracy have been used since the fourteenth century. A variety of direct democracy mechanisms are provided for at both federal and cantonal level, with Swiss voters given the chance to cast their votes in federal ballots on average four times a year.

When it comes to direct democracy, Switzerland is usually cited as the country that is the closest to having a directly democratic system of government. Although Switzerland retains features of a representative democracy (e.g. it has an elected Parliament), various forms of direct democracy are used frequently at national, cantonal and local (commune) level. This case study focuses on the mechanisms and use of Switzerland's direct democracy.

History and background

Direct democracy has a long standing tradition in some of the Swiss cantons, going back as far as the fourteenth century. When Switzerland became a federal state in 1848, direct democracy instruments were introduced at the national level as well. The federal constitution introduced the principle of holding a mandatory referendum in order to change the constitution, as well as the popular initiative for a total revision of the constitution. Further rights of referendums were introduced in 1874, and the popular initiative for a partial revision of the constitution in 1891. Between 1848 and February 2004, 517 referendums were held, whilst between 1892 and May 2004, 244 initiatives were proposed.

Forms of direct democracy - federal level

Numerous different direct democracy mechanisms can be used at federal level in Switzerland. The mechanisms fall into two broad categories: referendums and initiatives - there is no provision for use of the recall in Switzerland. Each mechanism can be used to achieve different results, and has different design features.

Referendums

Unlike in other countries, in Switzerland it is not the government that decides if a referendum is held on an issue; the circumstances under which referendums are used are clearly prescribed within the country's constitution.

The first type of direct democracy mechanism is the mandatory referendum, i.e., a referendum that the government must call in relation to certain important political issues. These are:

A partial or total revision of the federal constitution;
Joining an organisation for collective security or a supranational organisation;
Introducing urgent federal legislation whose validity exceeds one year, without the required constitutional basis (such legislation has to be submitted to the vote within one year after its adoption by Parliament);
Popular initiatives for a total revision of the constitution;
Popular initiatives for a partial revision of the constitution in the form of a general proposition which were rejected by the Parliament;
The question of whether a total revision of the constitution should be carried out if both chambers of Parliament disagree.

The first three kinds of mandatory referendums require a double majority to pass; that is, they must achieve a popular majority (a majority of the votes cast at the referendum) whilst at the same time achieving a majority vote in a majority of the cantons. The latter three, which take place as part of the initiative process, only need a popular majority.

Optional referendums can be held in relation to new or amended federal acts and/or international treaties. The optional legislative referendum is held in relation to all federal laws and urgent federal laws which are due to be valid for more than a year. The optional referendum on international treaties is held in relation to international treaties that are of unlimited duration and may not be terminated, and international treaties that provide for membership of international organisations or contain legislative provisions that have to be implemented by enacting federal laws Optional referendums are called if 50,000 signatures are collected in support of a referendum within 100 days, or if eight cantons request a referendum, and pass with a popular majority. Until 2004, an optional referendum has never been successfully requested by a group of cantons; the first referendum initiated by the cantons was held on 16 May 2004.

Initiatives

Initiatives can be used to propose changes to the federal constitution. In addition, in 2003 Switzerland adopted a new form of initiative, to be used in relation to more general statutory provisions. Once an initiative is filed, a specified number of valid signatures (i.e. signatures of registered voters) are required in order to force the Federal Council and Parliament to consider the initiative and to hold a referendum on the initiative proposal.

Amendments to the constitution can be proposed using two different initiative mechanisms. The popular initiative for a partial revision of the constitution provides voters with the opportunity to propose a draft revision to part of the federal constitution. 100,000 voters must sign an initiative in order for a referendum to be held on the proposal. The popular initiative for a total revision of the constitution also requires the support of 100,000 voters in an initiative. In both cases, the signatures must be collected within 18 months of the initiative being filed.

From late 2006, the general popular initiative has been available to Swiss voters. This mechanism can be used to force a referendum on the adoption of a general proposal that will be incorporated on a constitutional and/or legislative level, providing that 100,000 signatures are collected in support of the initiative.

Until 2006, initiatives in Switzerland were submitted as a general proposition or in the text that would be adopted if the initiative measure is successful. However, after the implementation of the general popular initiative, the popular initiative for a partial revision of the constitution will only be accepted in the form of a written text proposition (general propositions in relation to the constitution should be made using the general popular initiative). In response to initiatives which meet the required signature threshold, the Swiss Parliament advises the people on whether to adopt or reject the proposal. In addition, the government is also able to formulate a counter-proposal that is included on the ballot. The "double-yes" vote allows voters to approve both the original initiative and the government's response to it, and indicate which of the two measures they prefer. The measure which receives the most support is passed.

Forms of direct democracy - cantonal level

Use of direct democracy is even more extensive in Switzerland's 26 cantons (i.e., state authorities). However, use of direct democracy varies between the cantons; between 1970-2003 Zurich held 457, whilst Ticino held just 53 (the canton of Jura held just 45 referendums, but was only formally established (by referendum) in 1979).

In addition to the referendum and initiative mechanisms used at federal level, the following mechanisms are also used in some or all of the Swiss cantons.

Unlike at federal level, the legislative initiative has for some time provided voters in all cantons with the opportunity to propose additions to laws. In some cantons, the administrative initiative to launch a canton initiative can be used to demand that certain work is undertaken in public administration (e.g., building a new school or a new road). In addition, some cantons provide for the , an initiative to force the canton to table a motion to the Federal Assembly.

All the Swiss cantons provide for legislative referendums on legislation passed by the cantonal parliament; however, in different cantons, these may be mandatory or optional. Administrative referendums may be held on major public projects that will incur high levels of public expenditure (and may lead to increases in taxes); these are sometime called fiscal referendums. Lastly, administrative referendums may be held on the non-fiscal issues of public administration listed above.

Characteristics of the use of direct democracy in Switzerland

Turnout

Swiss voters are given the opportunity to vote in federal referendums on average four times a year. Typically, voters will also vote on a number of cantonal and local issues on the day of a federal ballot. Over the second part of the twentieth century, turnout at federal referendums fell from around 50-70% to an average of around 40%; this mirrored a similar decline in turnout at federal elections from 80% to around 45%. One suggestion is that this comparatively low turnout is due to the sheer number of votes that the Swiss are able to vote in; however, it is argued by many that a far higher proportion of the population is politically active than appears so from the figure of 40%, since it is not always the same 40-45% of voters who vote at each opportunity.

Issues

Given the numerous opportunities for using direct democracy in Switzerland, it is perhaps not surprising that the variety of issues on which referendums are held is extremely wide. Since 1990, referendums have been held on such diverse issues as:

Banning the building of nuclear power stations;
Building new Alpine railways;
A new federal constitution;
Controlling immigration;
Abolishing the army;
Joining the United Nations;
Shortening working hours;
Opening up electricity markets.

Impact of direct democracy

Undoubtedly, direct democracy has played a key role in shaping the modern Swiss political system. Yet it is important to question the actual impact of direct democracy on the legislative issues that, in other countries, are the responsibility of elected representatives.

On one reading, it could be argued that the impact has been limited: in the first century of using the initiative (1891-2004), just 14 initiatives were passed in Switzerland. Yet to consider this statistic alone ignores the considerable, indirect impact of direct democracy. Although the majority of initiatives fail, the fact that there has been an initiative, and therefore a campaign, increases publicity surrounding the issue in question and public knowledge of it. This may well increase pressure on the government to introduce measures dealing with the issue, even if it is not required to by virtue of a successful referendum. An initiative might therefore be successful in achieving some of its proponents' aims, even if it is not successful in the sense of having passed. This trend explains why many initiatives are filed but subsequently withdrawn; because sometimes a government chooses to act before an initiative reaches the referendum stage.

A further impact of the direct democracy mechanisms within Switzerland is that the government is forced to seek a wider consensus about the statutory (and constitutional) measures that it seeks to introduce than is the case in a purely representative system. In a representative system, the party of government may, in the absence of a large majority, have to develop cross-party consensus on an issue in order to ensure that the measure is approved. In the Swiss system, the possibility of an optional referendum forces the government to ensure consensus with groups outside of Parliament so as to prevent the possibility of such groups seeking to overturn the new legislation.

Conversely, the significance of direct democracy in the Swiss system is often cited as the reason for the weakness of Swiss political parties and the relatively low significance attached to normal elections. This is because, given the prominence of direct democracy, political parties are not solely responsible for controlling the federal agenda. In addition, direct democracy often raises cross-cutting issues on which members of political parties might not be in agreement.

Chile: A System Frozen by Elite Interests

Chile’s electoral system can only be understood in the context of the long period of authoritarian rule under General Augusto Pinochet (1973–90), whose aim was to establish a regime of protected, authoritarian democracy, of which the electoral system was one component. The dictatorship abolished PR, which had been in force prior to the military coup of 11 September 1973. PR was the response to the cleavages in Chile’s social structure since the 19th century and had produced a multiparty system. By the 1960s this had consolidated into six major parties—two on the left (the Socialists and the Communists), two in the centre (the Christian Democrats and the Radicals), and two on the right (the Liberals and the Conservatives, who merged in 1966 to form the National Party).

The Binomial System: a Legacy of Authoritarianism

In Chile’s bicameral constitutional arrangements, the Chamber of Deputies, the lower house, consists of 120 members elected for a four-year term, two for each of the 60 electoral districts. The Senate has 38 elected members, two for each of the 19 districts, elected for an eight-year term: there are elections for half of the seats every four years, simultaneously with elections to the Chamber of Deputies. There are in addition nine non-elected members, the ‘institutional’ or ‘designated’ senators, named by the National Security Council (four), the Supreme Court (three) and the president (two), and one ex-officio life member, former President Eduardo Frei Ruiz-Tagle. (The original 13 senatorial districts of the 1980 constitution were expanded to 19 in the 1989 constitutional reforms to reduce the power of the non-elected senators.) These arrangements were negotiated by Pinochet and his supporters as they fell from power during the transition to democracy.

Parties, coalitions or independents present lists, normally containing a maximum of two candidates per district, in elections both for the Chamber of Deputies and for the Senate. Voters vote for the candidate of their choice. The first seat goes to whichever list receives the most votes in total: the representative elected is the individual candidate on that list who receives the highest vote. To take both seats, the most successful list must receive twice the number of votes of the second list. This system forces the parties to form electoral coalitions because the effective threshold is very high: 33.4 per cent of the total vote for the top list is required to win one seat. However, a list needs to receive 66.7 per cent of the total vote to be guaranteed both seats.

There are two major electoral coalitions, which in 2001 won all the seats in the Chamber of Deputies except one. The centre–left Concertación por la Democracia is formed by four parties opposed to the Pinochet regime (the Socialists, the Democracy Party, the Christian Democrats and the Radicals) and ruled from the return to democracy in March 1990 up until March 2010. The right-wing opposition Alliance for Chile (the Independent Democrat Union, UDI, and National Renewal, RN) supported the Pinochet regime. In practice the Concertación list contains one candidate from each of two groupings within the coalition, that is, one from the Christian Democrats and another from the Socialists, the Democracy Party and the Radicals. There is no district in which there is competition between the Socialists and the Democracy Party. On the opposition list, the UDI and National Renewal normally present one candidate each in all districts.

The result of this electoral system is that almost all districts return one representative from the Concertación and one from the Alliance for Chile. The system could create competition between the two candidates on a list for the one seat it will win, but in practice even this is severely limited by elite accommodation within both coalitions.

This electoral system is unique because in practice it favours the largest minority, not the majority. It is thus not a majoritarian system. It is a system which uses a proportional mechanism, but the results it produces are not proportional, since it allows an electoral list to take half the seats with only 34 per cent of the votes. The only reason why this distortion has not occurred in practice is the limits to electoral competition.

The electoral system was set up by the military regime following the plebiscite of 5 October 1988. The plebiscite had two goals: to approve the 1980 constitution and to elect General Pinochet as president for a further eight years. In this non-competitive election (there was no other candidate), Pinochet was defeated by the Concertación. This triggered the transition to democracy, with congressional and presidential elections in 1989, the presidential election being won by the opposition candidate Patricio Aylwin (Christian Democrat). The electoral system was designed to favour the two right-wing parties, which had backed Pinochet’s candidacy, in the face of a predictable electoral victory for their opponents.

In the three presidential and four congressional elections held between 1990 and 2000, the Concertación has received most votes, but has never controlled the Senate because the majority of the institutional senators have supported the opposition.

The Drawbacks of the Binomial System for the Parties and for Democracy

Several objections to the electoral system have been voiced. First, it forces the parties into electoral coalitions because of the high vote threshold required to win a seat. Second, it has a negative impact on representation because it has kept the Communist Party out of Congress, despite its relevance up to 1973 and its 5–7 per cent share of the national vote in the new democracy. Third, since each coalition will normally win one seat, the real contest takes place among the member parties, rather than between rival alliances and parties. These disputes endanger stability in the coalitions; in the 2001 senatorial elections the UDI and the RN avoided them and named a single consensus candidate in seven of the nine districts, or ran only a weak competitor who would not challenge the leadership’s candidate. Fourth, the system hands enormous power to the party leaders, who virtually choose the winners when they make up the lists. With no real competition in many districts, the elections hold little interest for the voters, and even less so when there is no candidate of their own party to vote for.

The deficiencies have led the government to propose that there should be electoral reforms and to suggest that, instead of the two-member districts, larger districts that would yield more proportional results would be more appropriate. This has made little headway, however, because the Concertación parties fear the resulting uncertainty, and the opposition defends the current system because of the advantage it gives them.

Presidential Elections

The 1980 constitution establishes a two-round system for presidential elections. An absolute majority is required for victory in the first round, with a run-off round (ballotage) if this does not occur. The institution of ballotage tends to strengthen coalition politics. The winners of the presidential elections in 1989 and 1993—Christian Democrats Patricio Aylwin and Eduardo Frei, respectively—were elected with absolute majorities, but in 1999 there was only a scant 30,000-vote difference between Ricardo Lagos and his right-wing opponent, Joaquín Lavín. Lagos won with 50.27 per cent of the vote in the second round. (Under the previous (1925) constitution, when no candidate won an absolute majority, Congress decided the presidency, as occurred in 1946, 1958 and 1970. In each case it elected the candidate with the highest vote.)

Registration and Voting: Voluntary or Compulsory?

A further problem perceived in the current electoral system is that registration is voluntary but voting is compulsory. New electoral registers were opened in February 1987, when the military regime was preparing the October 1988 plebiscite, the old registers having been burned by the military in 1973. The democratic opposition mobilized strongly to get voters registered; its strategy was to defeat Pinochet at the polls in order to achieve democracy, and it succeeded in getting 92 per cent of eligible voters to register. Since then, however, the number of registered voters has not increased in line with the voting age population, as young people now show little interest in participating in elections. In the 2001 congressional elections 80 per cent of 10 million potential voters were registered; in the 2004 municipal elections the figure was 77 per cent.

Low registration among young voters led the government to propose automatic registration and voluntary voting. The Concertación parties support automatic registration, but there is no consensus on voluntary voting. They fear that overall participation will fall and that the financial costs of campaigning to mobilize voters will rise and rise, thus favouring the right-wing parties. The opposition, particularly the UDI, rejects automatic registration and supports voluntary voting.

Supporters of the binomial system claim that it has helped governability because there are two big coalitions, one in government and one in opposition. However, this view is mistaken: the Concertación as a coalition was created before the binomial system was introduced, as an alliance to work against authoritarian rule and promote a return to democracy by politicians who had learned from their past conflicts (which led to the crisis and breakdown of democracy in 1973) and had agreed on a strategy of elite cooperation within a political system somewhat comparable to a consociational democracy. The country is governable despite the binomial system, not because of it.

The system cannot last indefinitely because it damages the political parties and poses limitations to democracy, but it will be difficult to abolish because change would create uncertainty about the impact on party support. It would also require a constitutional amendment, because the binomial character of the Senate is in the constitution. There is consensus in Congress between the Concertación and the Alliance for Chile on eliminating the non-elected senators and former presidents as life members.

China: Village Committee Elections: First Steps on a Long March?

With the ‘household contract responsibility system’ introduced in the Chinese countryside in the late 1970s, farmers began to produce for their families. As production was decentralized, the collective-oriented organization of the People’s Communes became outdated.

The earliest villagers’ committees (VCs) emerged in the Guangxi Autonomous Region in 1980–1. Formed without the knowledge of the local authorities, these organizations were created by village elders, former cadres and community-minded villagers. The intention was to address a decline in social order and a broader political crisis as production brigades and teams stopped functioning at the grass-roots level. Within months, local officials had reported this development to the central government. The National People’s Congress (NPC) leaders encouraged experiments with this new form of organization.

In 1982, VCs were written into the constitution as elected mass organizations of self-government (article 111). In contrast to the relationship between the commune and production brigade or production team, the newly restored township—the lowest level of government—does not lead the VC but only exercises guidance over it. Another difference is the introduction of direct election by all eligible voters. In 1987, the Provisional Organic Law of Villagers’ Committees was passed, setting out general principles for direct elections to VCs and defining the tasks and responsibilities of the VCs. Implementation of the law, including the enactment of detailed regulations, was left to the provincial and lower-level authorities. The quality of elections and overall implementation varied considerably, and after ten years perhaps only 25 per cent of the more than 658,000 villages (for the end of 2002) in China had experienced direct elections in full accordance with the law.

In 1998, the NPC made the Organic Law permanent. The law has clarified and improved some aspects of the prescribed election procedures and strengthened the rules on transparency and popular control of VCs. The permanent law is seen by many as a political and legal consolidation of the village election process, but its full implementation remains a challenge—perhaps even more so after the introduction of more demanding standards, for instance, in relation to secret polling booths and the direct nomination of candidates. The quality of elections across the country still varies considerably.

The VC members are elected for three years, with no limit on the number of terms for which a person can be re-elected. The VCs usually consist of between three and seven members, one of whom is chair and one or two vice-chairs. Although there is variation from province to province, VCs generally oversee all the administrative matters of a village, including budget management, public utilities, dispute resolution, public safety, social order and security, health issues and local business management. A large village can consist of more than 10,000 people, while small ones might only have several hundred. The ‘average village’ has 1,000–2,000 inhabitants.

VCs report to the Village Assembly or the Village Representative Assembly. As the former meet only once or twice a year, the latter, composed of 25–50 people from the village and selected by Villagers’ Small Groups, play a greater role in decision making and in the supervision of the VCs. A Village Election Committee administers village elections.

Village elections have now been held in all 31 provinces, autonomous regions and municipalities. By 2003, the provinces of Fujian and Liaoning, two front-runners in this regard, had completed eight and seven elections, respectively, and 19 provinces had held between four and six elections. At least one province held its first village elections as late as 2000. There is no single election day for all VC elections across the country. During a province’s designated election year, the counties and townships within the province together decide the election days for the villages within their jurisdictions.

Each election adheres to the same basic framework. The first step in the process is the registration of voters, which is handled by the Village Election Committee. A list of registered voters must be prepared and publicly displayed 20 days prior to the election. Voters are allowed to challenge the registration lists.

Except for those who have been deprived of political rights, all those aged 18 or above enjoy the right to vote and to be elected without regard to ethnicity, race, sex, profession, family background, religious belief, level of education, property or period of residence in the community. One important challenge is the large number of voters whose residence registrations are in their ‘home village’, but who live and work a long distance away, often in a major urban area. It is difficult or impossible for most such voters to get back to their village on election day. At the same time, they cannot attend the elections in the cities in which they work and reside. Therefore they cannot actually exercise their right to vote.

Following voter registration, candidates are nominated directly by villagers. In most provinces, the requirement is to have only one more candidate than there are seats to be filled as chair, deputy chair, and ordinary members. In recent years, nominations in some provinces have been organized through villagers attending either a meeting of the Village Assembly or a meeting of the Villagers’ Small Group, while the latest development in other provinces is to have no pre-election nomination. In these areas, voters receive either a blank piece of paper or a blank ballot paper with only the different positions indicated above the relevant columns. If the election fails to produce a new committee or to fill all positions it de facto becomes a first-round election, and a run-off election follows.

The final election must be direct. The use of secret ballots and polling booths (or rooms) is mandatory in most provinces. There are three voting styles: (a) mass voting, where all voters go to a central voting place in the morning, vote, and remain there until the end of the count; (b) individual voting throughout the course of the day of the election; and (c) proxy or absentee voting, or ‘roving boxes’. Most of the provinces use mass voting. The ballot papers used contain names of candidates listed under the post for which they are standing; and the voting is done by the voter marking the names of the candidates he or she wishes to elect. The voter can mark as many candidates as there are posts (one chair, one or two vice-chairs, and a number of committee members) in the village. For an election to be considered valid an absolute majority of eligible voters must cast their ballots and winning candidates are required to get 50 per cent of the vote plus one. When no candidate receives a majority, a run-off election is held within three days. In run-off elections, candidates are only required to receive 33 per cent of all votes cast. Winners take up their positions immediately.

Village elections are important in that the election law mandates the basic norms of a democratic process—secret ballot, direct election and multiple candidates (even though their numbers are very restricted). Other elections in China have yet to implement these norms. The progress made in relation to VC elections has raised expectations as to whether and when direct elections will work their way up from the village to the township, county, and even higher levels of government. Each round of VC elections also strengthens local capacity to administer electoral processes.

An assessment of the significance of China’s village elections has much to do with the question whether such ‘limited democracy’ can lead to genuine democracy. There are different ways of assessing how democratic elections are. The three universal criteria of free, fair and meaningful elections are appropriate terms of reference. China does not meet any recognized standards of free and fair elections in choosing its national parliament and local councils, and in many cases elected village leaders do not exercise as much authority as the Chinese Communist Party (CCP) secretaries do. However, just because the village elections are not fully free or fair, and some VCs do not command complete authority, it cannot be concluded that they are completely unfree, unfair, or meaningless. Elections should not be evaluated against some absolute standard but rather viewed as positioned on a democratic continuum.

The VC elections have produced a ripple effect as village CCP branch elections in some cases have invited ordinary villagers to cast a vote of confidence, and some experiments with elections of township government leaders have taken place. China’s democratization now appears to require that the top leadership’s political decisions find an echo at the grass roots. After two decades of continuously improved direct elections at the village level, elections at higher levels of government appear technically feasible; the question is whether and how there will be further change in the direction of democratization.

Germany: The Original Mixed Member Proportional System

After the use of the absolute-majority Two Round System (TRS) in the German Empire, and the use of a pure proportional representation system in the Weimar Republic, see, a new electoral system was established by the Parliamentary Council in 1949. The system was created by the Basic Law of the Federal Republic of Germany - the West German Constitution. It was thus a result of inter-party bargaining between democratic forces in West Germany. Like the Basic Law, it was originally considered to be provisional, but has remained essentially unchanged since 1949.

The German electoral system is classified as a personalised proportional system ("Personalisierte Verhältniswahl") or, as it is known in New Zealand as a Mixed Member Proportional (MMP) system. Its essence is the way in which it combines a personal vote in single-member districts with the principle of proportional representation.

Currently, the German parliament (Bundestag) has 622 seats, not including possible surplus seats (see below). Each voter has two votes. The first vote (Erststimme) is a personal vote, given to a particular (party) candidate in one of the 299 single-member constituencies. The second vote (Zweitstimme) is a party vote, given to a party list at the federal state level (Landesliste). Candidates are allowed to compete in single-member districts as well as simultaneously for the party list. The candidates who achieve a plurality in the single-member districts are elected (Direktmandate). However, the second vote determines how many representatives will be sent from each party to the Bundestag.

On the national level, all the second votes (Zweitstimmen) for the parties are totalled. Only parties obtaining more than five percent of the votes at the national level or, alternatively, having three members elected directly in the single-member constituencies, are considered in the national allocation of list PR seats. The number of representatives from each party that has passed the legal threshold is calculated according to the Hare formula. Seats are then allocated within the 16 federal states (Länder).

The number of seats won directly by a party in the single-member districts of a particular federal state are then subtracted from the total number of seats allocated to that party's list. The remaining seats are assigned to the closed party list. Should a party win more Direktmandate seats in a particular federal state than the number of seats allocated to it by the second votes, these surplus seats (Überhangmandate) are kept by that party. In such a case, the total number of seats in the Bundestag temporarily increases.

The German system is not, as sometimes supposed, a mixed system, but a PR system. It differs from pure proportional representation only in that the five percent threshold at national level excludes very small parties from parliamentary representation, and thanks to proportional representation a relatively wide range of social and political forces are represented in Parliament. Furthermore, the electoral system is, to some extent, open to social and political changes. In spite of the threshold, new political parties supported by a substantial part of the electorate have access to Parliament. Besides the Christian Democratic Union/Christian Social Union (CDU/CSU), Social Democratic Party (SPD) and Free Democratic Party (FDP), which have been in the Bundestag since 1949, a new Green Party (GRÜNE) gained seats in 1983 and 1987. After falling below the threshold in 1990, the Greens, in a coalition with Alliance '90, were able to return to Parliament in 1994. After German unification, even small East German parties gained parliamentary seats. In the all-German elections of 1990, the East German Alliance '90/Greens and the Party of Democratic Socialism (PDS) cleared the five percent threshold which was applied, separately in the territory of former East Germany and former West Germany, for that one election. Four years later, the PDS took advantage of the "alternative clause" by winning four of the required three Direktmandate seats.

The personal vote for a candidate in single-member constituencies aims to ensure a close relationship between voters and their representatives. In practice, however, the advantage of these districts should not be overestimated. In Germany, elections in the single-member districts are mainly based on party preferences and not on the personality of the candidates. The initial hopes that MMP would guarantee a close voter-representative relationship have consequently only partly materialized, despite efforts by representatives to establish strong links with their constituencies. Nevertheless, this constituency element within a PR system does at least help to bridge the gap between voters and representatives which is normally widened by ordinary closed-list PR systems.

Furthermore, the two-vote system enables voters to split their votes strategically between existing or possible coalition partners. In fact, vote splitting is common among the supporters of smaller parties. Since candidates of smaller parties have little chance of winning a single-member district, their supporters frequently give their first vote to a constituency candidate from the larger coalition party. Similarly, supporters of bigger parties may "lend" their second vote to a minor party within the coalition, in order to ensure that it will pass the legal threshold. Thus, vote splitting is strategically used by voters to support the coalition partner of "their" party or, at least, to indicate their coalition preferences.

By producing highly proportional outcomes, the electoral system makes manufactured majorities, where one party wins an absolute majority of the parliamentary seats on a minority of the popular votes, very unlikely. In fact, over the last five decades in Germany, manufactured majorities have never occurred. Majority governments have usually been coalition governments, and any change of government has resulted from changes in the configuration of the coalition. German coalition governments are usually stable and regarded as legitimate by the electorate, and, because of a coalition's built-in incentives to co-operate, many Germans prefer a coalition government to a single-party government. The main checking function is fulfilled by an opposition, which is fairly represented. It is important to note that the relationship between government and opposition in German politics is more consensual and co-operative than conflictual or hostile. This, however, is a result of history and political culture rather than of the electoral system per se.

To date, the MMP system has not shown any great drawbacks in Germany. It has lasted long enough to have a high level of institutionalized legitimacy; the basic principles of single-member districts and list PR representation have been left unaltered since 1949. However, some minor changes of the electoral system have taken place. Chief among these was the switch to two separate votes in 1953, before then the voter had only a single vote to apply to both district and national PR allocation.

Nevertheless, several attempts to reform the electoral system substantially have been made since 1949, and most intensely in the 1960s, when opponents of the PR system demanded the introduction of a FPTP system. This was partly due to political manoeuverings to enhance the position of the stronger parties, and partly based on a theoretical school of thought which favoured the British model; but all attempts were unsuccessful. More recently, the electoral system has been criticized for producing too many surplus seats without compensating the disadvantaged parties in Parliament.

Ecuador: The Search for Democratic Governance

Aware of the constitutional and methodological inconsistencies of the previous electoral rule, Congress approved a reform of the Electoral Law in March of 2000 to introduce "vote pooling". In this modified version, each party pools votes obtained by individual candidates and the total sum of votes obtained will be used to distribute seats according to a D'Hondt formula of proportional representation. Party leaders in turn distribute allocated seats to the most voted candidates within each party. The revised version of the electoral formula will be first applied at the national level in the 2002 general elections. Finally, the election of National deputies was also eliminated for the 2002 election, leaving Congress with provincial representations only.

Despite extreme and frequent experimentation with its electoral framework, Ecuador has not been able to promote an effective combination of representative politicians with effective government. Over time, there has been a permanent government effort to manipulate the electoral system to reduce party system fragmentation and encourage the formation of pre electoral alliances that could eventually lend support to government-led initiatives in the legislature. The adoption of a two round system of presidential election and its modified threshold or the overrepresenting of larger parties through a modified PR formula were meant to give both legitimacy as well as ample party support for presidential candidates when taking office. Opposed to the logic of "majority building", different social and ethnic groups, as well as entrepreneurial politicians claimed that the electoral system privileged party interests over citizens' demands and sought to relax electoral restrictions to political participation and promoted the candidacies of "independent" politicians. In turn, the proliferation of independent candidates in the mid and late 90's made coalition formation in Congress unpredictable.

The constitutional battle over the requisites to maintain party registry well illustrates the tension between governability and representation. The 1979 Law of Political Parties established that parties that did not obtain 5% of effective votes in two consecutive elections would lose their electoral registry. This prohibition was declared unconstitutional in 1983, reinstated in 1994 as 4%, abolished in 1996 and reinstated in 1998 as 5% again. While the advocates of governability argued that smaller parties contributed to legislative fragmentation and unpredictable coalition building in Congress, defenders of representation argued that Ecuadorian minorities deserved to be represented by those parties. In the meantime, small and usually personalistic parties were able to survive for several years.

Ecuador has not found a power sharing formula to promote democratic governance. One open issue for debate is the adoption of a mixed-member system for the legislature. In principle, this could reconcile the need for direct provincial representation of the heterogeneous Ecuadorian population with the election of nationally oriented politicians. Another issue is the introduction of some parliamentary features like cabinet sharing would make potential coalition partners (parties) more accountable to their electorate and responsible to the government. In any case, an effective electoral system would need to mature over time (without being subject to sudden changes) and would have to take into consideration other political institutions and historical traditions of the country.

European Parliament: Elections to a Supranational Body

In June 2004 the citizens of 25 European countries went to the polls to elect their representatives to the European Parliament. Many were doing so for the first time, while others were in countries with experience of up to five previous elections to the Parliament. In 2004, all were conducting their elections under a proportional electoral system.

The first piece of legislation covering elections to the European Parliament came in 1976, when the Act Concerning the Election of the Representatives of the Assembly by Direct Universal Suffrage was agreed. As the name implies, this act determined the principles for direct elections of the representatives from each member state. In the early days of the EU, the members of the European Parliament were nominated by the legislature in each member state, with no direct input from the electorate. The passing of the 1976 act meant that from then on the members would be elected by direct universal suffrage in each member state; and in 1979 the first European Parliament elections were held in the then nine member states—Belgium, Denmark, France, the Federal Republic of Germany, Italy, Luxembourg, the Netherlands, the Republic of Ireland and the United Kingdom—resulting in a total of 410 elected members.

The act of 1976 determined many things regarding the elections, such as the length of the parliamentary term and the eligibility of candidates, but did not in itself determine the actual electoral system to be used in these elections. It did, however, give the European Parliament the task of drawing up a proposal for a uniform electoral procedure. Until such a procedure came into force, the act left the electoral procedures to the national provisions of the member states. As most member states at the time were using a PR system of one form or another to elect their legislatures, either alone or as one component of the electoral system, the choice of which electoral system to use for the European Parliament was a simple one. Belgium, Denmark, the Federal Republic of Germany, Italy, Luxembourg and the Netherlands were already familiar with the List PR system in one form or another, and all their representatives were therefore elected under a List PR system (except for the one representative of Greenland, included within the representation of Denmark, who was elected by FPTP until Greenland left the European Community in 1985). The Republic of Ireland chose to elect its representatives using its preferential STV system.

There were only two exceptions: the UK with its FPTP electoral system and France with its Two-Round system for national elections were both unfamiliar with the PR system. The UK simply copied the electoral system used for the elections to the House of Commons and applied it to the European Parliament elections as well. This system suited the two largest parties, Labour and the Conservatives, very well, and made it difficult for any third party to enter the arena. Resistance against a representative holding a dual mandate also contributed to the adoption of FPTP for European Parliament elections in the UK. If serving in both the European Parliament and the British Parliament simultaneously were to be prohibited, as some British politicians wanted, and some form of PR were also to be adopted, party lists would be likely to be made up of unknown candidates not elected to any other national or local body, which would carry the risk of undermining the perceived importance of the elections. A candidate-centred, single-member district system, where candidates would be closer to their electorate, was thought to be a better solution. None of this thinking applied to Northern Ireland. Concern to ensure the representation of majority and minority communities, combined with the fact that the parties of England, Scotland and Wales do not normally contest elections there, resulted in the use of STV for Northern Ireland’s three seats. An attempt was made to challenge in the courts the use of FPTP for European Parliament elections in England, Scotland and Wales on the basis of the requirement contained in the act for a uniform electoral procedure to be proposed, but this was unsuccessful. The system used in England, Scotland and Wales only changed in 1999 when the European Parliamentary Elections Act was passed, as the UK anticipated the changes that would be forced on it as the process which led to the 2002 Council decision (see below) got under way. From 1999 onwards, the UK joined the other member states in the use of a PR system, choosing List PR with closed lists and regional electoral districts.

In France—despite its using TRS for the elections to the national legislature—a closed List PR system with one national district was adopted as early as 1977, before the first European Parliament elections in 1979. The reasons for this were many. One of the main advantages of a plurality/majority system—the formation of stable majority governments—was clearly not relevant for these elections, and the proportional representation of all political parties was seen as a much more important criterion for the design of the electoral system. The nationwide district in combination with a 5 per cent threshold was thought of as providing a balance between a high level of proportionality on the one hand, and the desire to exclude parties with little support on the other. The ability to fill vacant seats between elections with the next person on the list, thus eliminating the need to hold by-elections, was another advantage that led to the adoption of the List PR system. After five elections, the wish to strengthen the relationship between voters and representatives, and the desire for greater geographical representation (a disproportionate number of those elected had been residents of Paris), led to the nationwide district being abandoned before the election in 2004. It was replaced by eight multi-member districts for the election of France’s 78 representatives to the European Parliament.

The next piece of important legislation on the European Parliament elections was concluded in 2002 with the Council Decision 8964/02 amending the Act Concerning the Election of the Representatives of the Assembly by Direct Universal Suffrage—an amendment to the 1976 act. Twenty-six years after the establishment of provisions for direct elections, this decision specified a common electoral system family for European Parliament elections for all member states. Article 1 reads: ‘In each Member State, members of the European Parliament shall be elected on the basis of proportional representation, using the list system or the single transferable vote.’ For the 2004 elections, all 25 member states thus used a PR electoral system.

While all these systems belong to the same family, they also differ in some respects. Twenty-three countries (Austria, Belgium, Cyprus, the Czech Republic, Denmark, Estonia, Finland, France, Germany, Greece, Hungary, Italy, Latvia, Lithuania, Luxembourg, the Netherlands, Poland, Portugal, Slovakia, Slovenia, Spain, Sweden and the UK) use List PR, while Malta and the Republic of Ireland use the Single Transferable Vote. This difference may be expected, since the List PR system makes up at least a part of the electoral system for the national legislatures of 21 of the 23 countries (France and the UK being the exceptions), while the Republic of Ireland and Malta both use STV for their national elections. In the 23 countries using List PR, some used closed lists, while others have chosen open lists—a choice which in most cases mirrors that made for national elections, although Greece is one exception. Equally, some member states, especially the smaller ones, elect their members from one national district, while others have set up a number of districts at regional level. The threshold for gaining representation in the different member states also varies. The 2002 decision allows for the individual countries to determine the threshold, but sets the ceiling for any formal threshold at 5 per cent. Some countries, for example Cyprus, Hungary and Sweden, use formal thresholds, again for the most part mirroring their use at national level (although Belgium, which uses a formal threshold for national elections, does not do so for European Parliament elections). Not only the formal thresholds, but also the actual level of support needed to gain representation—natural thresholds—vary significantly between member states. The reason for the variations in thresholds is found in the combination of the number of representatives to be elected from each country and the level at which the electoral districts are defined (more specifically, the number of representatives to be elected from each district). Italy, with one nationwide district and 78 representatives to elect, has a very low effective threshold of under 1 per cent, while the four electoral districts and 13 representatives of the Republic of Ireland mean that a successful candidate under STV will need to win a much higher proportion of the vote. In 2004, the winning candidates in Ireland received between 12.9 and 25.9 per cent of the first-preference votes in their district.

While a common electoral system family is now specified for European Parliament elections, there is little sign of any momentum for further integration. Although party groups are formed within the European Parliament, there is no sign that national parties are willing to relinquish any significant leading role to pan-European parties. There is thus every likelihood that decisions about electoral system details will remain in the hands of national politicians, influenced by their own interests and their existing national traditions. Debate seems more likely to centre on the low voter turnout in European Parliament elections, which remains a major concern of the EU member states. Despite the use of a PR system in all countries—an electoral system family which is usually linked to a higher voter turnout than other systems—turnout is still strikingly low. At the 2004 elections, the 15 countries that were members before enlargement in 2004 had an average turnout of 52.9 per cent, and the 10 new member states an even lower figure of 40.2 per cent. It appears that as long as the electorate sees European Parliament elections as being secondary, with little clarity as to what changes result when representation changes at elections, interest and turnout will remain low. The electoral systems used are not seen as a controversial element, and there is very little serious debate about their amendment. It is therefore likely that the electoral systems will remain fairly constant in the near future.

The Spanish Electoral System - Historical Accident

Historical Background

The electoral system in Spain has a long history. It began early in the 19th century, in extraordinary circumstances arising from the invasion of the country by the Napoleonic army. The power vacuum created opened the way for a call for a national assembly (las Cortes) by a Central Council (Junta) - the leaders of the resistance - in the Andalusian city-port of Cadiz. The deputies, who came from all regions of Metropolitan Spain and abroad, were elected by popular vote (according to the "Instructions for the election of deputies" of 1810, which was attached to the notice of meeting and was the embryo of electoral law). This assembly, at its first meeting, declared itself a sovereign and constitutive body and became the first modern parliament in Spanish history. It ended with the enactment of the Constitution of 1812. The text of the Constitution clearly drew its inspiration from liberal thinking (the term "liberal", applied to politics, has its origin internationally in the name that was given to the group of progressive deputies in the Cadiz parliament) and for Spanish people has always been the symbol of political freedom. The drafting of this first Constitution in the Spanish language, which incorporated many ideas from the French revolution and models from the Constitution of the United States, also found inspiration in the old democratic traditions of the Spanish municipalities. In the constitutional text of 1812, the latter regained their former status of elective institutions (the municipalities in the historical and political history of Spain have always played a key role in the defence of freedom).

The "Cortes de Cadiz" (the Cadiz Parliament) have been regarded as the starting-point for democratic ideas and the right to vote, not only in Spain but also in the whole of the Hispanic world. This was because, together with the deputies from Metropolitan Spain, those from the then Spanish provinces of America (one of the three groups of representatives who met in Cadiz called itself the party of the "Americans"), participated in drafting the text. Some of the latter would later be leaders of movements for independence in Latin America. The Constitution of Cadiz would thus inspire not only subsequent Spanish constitutions, but would also serve as the basis for the first constitutions of the new independent Latin American states. (We would say in passing that in almost all Latin American countries the process of independence began in the struggles for the democratization of power in the town councils.) It is interesting that, for this reason, many of the first electoral laws in the 19th century in Spanish-speaking countries on both sides of the Atlantic had as their common point of reference the same constitutional text.

Following from this essential introduction, the electoral laws that were to govern the right to vote began formally in Spain from 1837, with universal suffrage being established in 1869, and were consolidated by the law of 1907. However, Spain's stormy history throughout the 19th century and part of the 20th century meant resounding retreats and advances in the process towards the exercise of democratic liberties and hence towards the functioning of the electoral system.

The Second Republic, proclaimed in 1931 after the results of the municipal elections that forced the king to abdicate, also introduced women's suffrage. Spanish women, who voted from the General Elections of 1931 onwards, did so long before this happened in other countries of Western Europe (before France for example). The triumph of the Popular Front in the elections of 1936, the results of which were doubted by none, within a few months unleashed an anti-democratic reaction and the outbreak of the civil war from 1936-39. It ended in victory for General Franco made possible by the internecine conflicts of the republican forces and by the strong external military support from fascist regimes.

It is a known fact that the establishment of a dictatorship imposed for nearly 40 years an interruption in the exercise of democratic rights. Nevertheless, these nearly four decades were not sufficient, as the Spanish political transition later came to prove, to wipe either the past or the accumulated experience of a century of electoral legislation, from historic memory: between 1810 and 1936 twelve electoral laws had been passed and many other regulations governing the holding of 55 parliamentary elections in Spain over this long period.

The fact was, even before the death of Franco and within Spain, that in intellectual circles and among members of the democratic opposition parties who were beginning to come out of political hiding, the subject of the most appropriate electoral system for the country was already being discussed and studied. Immediately after the death of Franco in November 1975, this question leapt into the media and an extensive debate was begun on the advantages and disadvantages of both the electoral system operated during the Second Republic and of the different systems used in other countries.

Development Within The Political Transition

Following the death of the aged dictator, an almost general consensus agreed on the necessity of reaching a pact between the political forces, including the reformist wing of the old regime. A peaceful transition had to be made by means of just, transparent, and reliable elections through an electoral system which would give suitable opportunities to the whole of the wide political spectrum, including the nationalist parties, to compete for representation in the future parliament.

One year after Franco's death, in December 1976, the Spanish people were asked to approve by referendum the law for the Political Reform of the state. Avoiding any break with old institutions, this signified a first recognition of the principle of popular sovereignty, enabling citizens to declare themselves freely and to choose either a political system based on democracy, or a continuation of the dictatorship. In the first alternative, by voting "Yes", legitimacy and a free mandate would be given by the vote to the calling of a democratic constitutive parliament, which would have to be elected six months later. It would be based on an electoral system that would have to be approved should the result of the referendum be "Yes" to democracy. The results were overwhelming for the "Yes" vote, as only two-percent of the Spanish electorate voted "No" to democracy.

Approved by referendum, the political reform set up a bi-cameral parliament, comprising a Congress of 350 deputies (based on a ratio of 1 deputy per 100,000 inhabitants) and a Senate of 297 senators (this number having altered subsequently). The debate on the establishment of an electoral system to form these chambers revolved around the two principal aspects which are the foundation of all electoral systems: on what territorial basis should the constituencies be established; and which electoral formula was it appropriate to adopt in the historic/political context of the country?

The strong influence of history upon the various Spanish regions meant that the electoral system adopted as from 1977 had to balance the purely "population" component (i.e. had to assign to each territorial demarcation a number of deputies for election in accord with the number of inhabitants with the right to vote) with a formula which would allow the population of each territory to have a minimum representation in accord with the variable of territorial size. We should point out that Spain is a country characterized by great demographic imbalances over its land area. Hence, the electoral system adopted in regard to assignment of seats per constituency was based on a "two-tier" mixed system of proportional representation, combining elections at the provincial level with national party lists.

The territorial demarcation of each constituency was linked to the division of Spain into provinces (there are 50 provinces), to which were added two constituencies covering two Spanish cities situated outside the peninsula. The distribution of the 350 Congressional seats was made in such a way that each constituency would have two seats permanently assigned to it on a territorial basis, with the rest of the seats being distributed by assignment to each constituency in accord with the variable of "population". This is the variable which makes it possible that in some cases, from one election to another, the number of deputies which each constituency may elect can vary slightly, the Constitution subsequently determined the number of seats able to be held in the Congress of Deputies at between 300 and 400. In practice, however, the initial 350 parliamentary seats have been maintained to date, divided between the 52 constituencies according to the system described.

To compensate for the effects of assigning seats to the constituencies on the basis of this "territorial/population" system (which favoured some candidacies more than others assign), the electoral system sought a corrective element in the formula for turning votes into seats. The simple majority systems such as FPTP or TRS, which would have accentuated the disproportional effects of the "two-tier" mixed system, were ruled out, and the system of closed party-list PRs was chosen, with the d'Hondt formula used to allocate seats. Multiple ballot papers containing closed blocked lists and Hondt's Law on proportionality was instituted for the adjudication of seats per province, which could in this case favour other candidacies. At the same time, the exclusion barrier against a candidacy entering into the distribution of seats was established at a minimum of three percent of the vote in each constituency.

The system for the Senate, which is a chamber of territorial representation, currently composed of 264 senators (this can vary by one more or one less), is organized very differently, since only 208 seats are elected by direct election. These are assigned by the distribution of four seats to each of the provincial constituencies, independently of their population, the elector being able to vote for three of the candidates who appear on a single ballot paper (on which each political grouping presents three candidates for election and has three boxes). Under this system of "open list" PR for example, it is possible to choose to vote for a single candidate or for three candidates of three different parties. The remainders of the senators (56 in the current legislature) are elected by indirect vote by the parliaments of the 17 autonomous regions ("Communities").

With regard to active voting rights (to be able to elect) and passive voting rights (to be eligible), the Spanish electoral system gave both practically without restriction to all citizens of full age (18 years), excluding only those convicted by final sentence of the courts or, in the case of candidates for election, those who actively exercise certain public functions (judges, the military, high positions in the administration, etc.) Furthermore, candidates would not be obliged to make any kind of financial deposit to compete in the elections. It was considered preferable to eliminate any kind of discrimination in the submission of candidacies for financial reasons, even at the risk of some possible abuses, although they would have to be presented by a legally registered party (to register a political party is extremely easy) or by a group of electors. Finally, regarding participation in an election, the elector was left entirely free to decide whether to exercise his right to vote or not - the imposition of compulsory voting in the context of transition from a dictatorship to a democracy would have been an ideological contradiction.

According to the planned time-scale, three months after the constitutional referendum, the first electoral rules (as described above) were provisionally approved, and three months later the general elections to appoint the constitutive parliament took place. The integrative capacity of the electoral system described - fundamental for a political transition to be truly viable - despite imperfections such as are found in any electoral system, was proven by the fact that the right, the centre, the socialists, the communists, and the Basque and Catalan nationalists obtained a parliamentary representation sufficiently aligned to their expectations. This parliamentary plurality, obtained without excessive fragmentation and which reflected the great ideological currents within the country, was fundamental to a real consensus on the Constitution text. Its complex preparation required more than a year and there is no doubt that from the standpoint of comparative law, it can be considered from many aspects to be one of the most advanced in the world.

Among the provisions of the Constitution (Article 8), perhaps one of the most important and far-reaching was that of granting Parliament exclusive power to draw up the electoral rules, and also that of establishing that electoral law should have the status of constitutional law. Any amendment, however minuscule, should be submitted to the scrutiny of the Constitutional Commission of Parliament and follow the formal procedures reserved for constitutional laws.

After approval by referendum in 1978 of the new Spanish Constitution, the constitutive Parliament, which had fulfilled its function of drawing up the basic law of the State, was immediately dissolved. New General Elections were called which would pave the way for the first ordinary legislature, and also the first municipal elections of the democracy, so bringing to an end the initial phase of the Spanish political transition.

Legal And Functional Focus Of The Electoral System

The Spanish electoral system, as essentially an instrument for guaranteeing an egalitarian and democratic vote and translating votes in terms of political representation, has been a truly effective model, so much so that it has continued to remain virtually consolidated and unaltered for twenty years after its principal outlines were approved by consensus in the initial phase of transition. (It has made possible majorities and relatively stable governments, as well as the changeover in 1982 and in 1996.) The first Constitutional Electoral Law passed in 1985 ratified the bases of the system devised in 1977 and expanded the development of the regulations applied during the period of political transition.

It is quite true that a certain debate has been opened on the opportunity of introducing a few modifications into the electoral system and adopting the system of open lists instead of closed lists, with timid proposals being heard favouring the establishment of single-member constituencies and candidacies. However, it seems very unlikely that in the complex context of Spanish politics/elections, the advantages of other systems will compensate for the difficulties that would arise, especially regarding single-member constituencies, of which all political circles are aware.

This being said, the validity of the Spanish electoral system lies both in its being politically functional (within the framework of Spanish society) and in the legitimacy that it has gradually acquired throughout the numerous electoral processes organized in the thirty-year life span of the system. No political force of any persuasion has ever brought a global accusation of electoral fraud, and the irregularities complained of have always been very parochial. All candidates accept the provisional electoral results announced by the Home Office on the night of the ballot as reliable, by the media and by the electorate.

The reason for this was the deep-rooted conviction held by those who drafted Spanish electoral legislation, that any electoral system, which theoretically might be considered the best in the world, could in practice be useless. If the procedures for its application left any margin for manipulation in such essential areas as the compilation of electoral lists, the registration of candidacies, the counting of votes and many others, both technical and administrative, which directly affected the democratic validity of the electoral process.

To avoid risks of this kind, both in the Constitution and in the first draft of the Electoral Law (L.O.R.E.G.), Electoral Law was considered to be a branch of Constitutional Law. A legalistic approach was established, imposing maximum guarantees and restraining subjective interpretations facilitated by silence or ambiguities in the rules and preventing the executive power from dictating "ad hoc" rules which might lead to possible manipulations in any phase of the process of establishing the right to vote.

In Spain therefore, by virtue of this philosophy, the Government or the administration in its wider sense has no legal power to regulate or introduce provisions affecting the electoral regulations. The executive is merely able to approve a decree calling an election or technical decrees to be applied to the law. Ever since the Constitution was drafted, the executive and the administration have only played an instrumental role, and are merely organizers of the electoral process. This premise is now embedded in Spanish electoral culture, and although the electoral system has not changed fundamentally, the aspect that is developing through parliamentary debate is that of the actual functioning of electoral regulations in relation to the organization of the electoral process, its control, and technical modernization.

It is to be noted also that the concept of Spanish Electoral Law, since its adoption as constitutional/organic law, is as a unit and, as stated in its preamble, responds to "the need to treat in a unified and global manner" all aspects of the electoral process. There are no lack of examples in western Europe of models of the "puzzle" type of legislation, where the Law or Electoral Code is a general regulating framework which needs to be accompanied or completed either by other laws or by a body of laws, or else by decree, regulations, circulars, etc. emanating from the executive. (Instead of facilitating transparency these finish by creating a legal labyrinth). However, the Spanish parliament at that time ruled out the use of any outdated model of fragmented electoral legislation as a reference and decided to incorporate all provisions into the same text, so that it would be functional and coherent, and clearly and concisely written. In this sense, the Electoral Law constitutes a veritable manual of procedure, covering in ordered fashion not only the constitutional principles guiding the electoral system, but also the precise provisions and their time schedules which must regulate the control, financing, management, and administration of any electoral process, taking into account also their different classification (parliamentary, local etc. elections).

Description of the Functional Elements of the System

Operations of Supervision and Control of the Electoral Process:

The body that is charged with these functions has been created with the name of Electoral Council, which in some Latin American countries is called "electoral power." The Electoral Council has its own hierarchy and is based on territory: Central Electoral Council; Provincial and Area Councils; and Electoral Councils of the Autonomous Communities (in Spain there is a system of practically federal decentralization, with 17 autonomous regions). The composition of this Body is mixed one part of its members being appointed by the Council of Judicial Power and the other by Parliament.

The Central Electoral Council is permanent; it has 13 members of whom eight are judges of the Supreme Court appointed by lot and the other five members, appointed by Parliament, hold University Chairs in the fields of Law, Political Sciences, or Sociology. The President is elected from among the judges and the Secretary (without vote) is the General Secretary of the Congress of Deputies. The Central Electoral Council is renewed 90 days after the constitution of a new Parliament following the elections. Its office is within the Parliamentary buildings. Electoral Councils, within their respective territorial ambit, have absolute power over all matters relating to elections and to the organization of the process, although for reasons of efficiency they do not take on the specific tasks of physically organizing the elections, but devote their services entirely to the validation of the operations and to the control and supervision of their legality.

Operations of preparation and revision of the Electoral Census: Spanish citizens acquire the capacity to vote and their eligibility to do so on attaining 18 years of age, and are automatically included on the electoral lists. Spaniards ordinarily resident abroad are also included on special electoral lists. Foreigners of member countries of the European Union and of Norway, ordinarily resident in Spain, are also entered on the electoral lists, are able to vote and are eligible in municipal elections. Those who have attained 17 years of age are included on an attached list, so that they can vote if they have reached 18 years on polling day. No elector's card is issued since identity is checked at the electoral tables; this is by identity card or passport. Electors (currently some 30 million) receive a card at their home address, which is not valid for voting, but shows that they are registered on the Census, and the electoral table to which they belong. The management of the Electoral Census is in the hands of the Electoral Census Office, an organization which is under the direct supervision of the Electoral Councils and financially dependent on the National Statistics Council (attached to the Ministry of Finance). The basic information for the census is supplied by the City Councils (there are just over 8,000 municipalities which must compulsorily declare increases and decreases in population, changes of address, and changes in cartography), and by Consulates and Civil Registers in the case of deaths.

An important step in the modernization of the management of the Electoral Census has been taken recently with the reform of the Electoral Law of 1995; instead of the previous yearly updating of the electoral lists, a monthly updating has been imposed. The considerable investment in the relevant computerization placed upon the executive by the actual reform of the Law means that in the short term the margin of error is minimal. Currently, every election takes as its basis the electoral census of the month before it was called, and the lists for each electoral month are exhibited publicly at the polling stations one-week after the call is made. Electors can object immediately in cases of error or non-registration.

Operations of organizations relating to the electoral process: The Home Office is responsible for the logistics and financial management of expenses incurred in the organization of elections. It must also, of course, attend to security for the peaceful running of the electoral campaign and the free exercise of the right to vote. The various police and Civil Guard corps receive training in their academies on these matters and on polling day carry a pocket card which reminds them of the relevant provisions of the Constitution and the Electoral Law. Regarding logistics and financial management, a Deputy Manager is charged with organizing the elections and planning the electoral time schedule in accordance with the provisions of the various articles of the Electoral Law which, as has been said, are very precise and determine all time periods, including the hours of opening and closing of the polling stations.

In the areas of electoral management within its remit, the Home Office has a dependent relationship vis-a-vis the Central Electoral Council, and always consults the latter on any problem or query that may arise in the various phases of the electoral process. The instructions of the Electoral Council - even if sometimes the election technical workers have different opinions when their practical application poses problems - are heeded without delay.

At territorial level, on the instructions of the Home Office, the State administrative services in each province take charge of the specific tasks of storage and distribution of electoral material, as well as of the printing of ballot papers and envelopes. The Home Office also contracts with publicity agents on audiovisual campaigns which are broadcast during the process to inform or remind electors of certain matters (display of lists, procedure for voting by post, identity papers required to vote, etc.).

A very important function of the Home Office is that of providing the public with the provisional results of the vote-count, to which enormous human resources are devoted. Once the counting of votes is over, thousands of electoral agents representing the administration take a copy of the results at each one of approximately 50,000 electoral tables distributed throughout the country, and send them by telephone to regional computerized centers, which process the information and send it to the central computer. In Spain the speed of transmission of the provisional results is greatly assisted by the fact that all the electoral colleges have a telephone on the premises or close by. Considerable funds are also devoted to this operation (a figure of some US$5,000,000 could be indicated.)

The result of all this great effort on the night of the election has been to make ever shorter the time taken to inform the country of the detailed results of the elections, so that four hours after the close of the voting operations at the 50,000 electoral tables, the detailed count is computerized and made known practically one hundred percent. The competing political groupings and the communication media have direct lines connected to the central computer at the Home Office, giving them access in real time and from the beginning, to the development of the vote count. From 1996 the counting process has been made known worldwide on the Internet. Currently, the feasibility of introducing the electronic vote is being studied, although it is not yet very clear if the investment in this new technology - some testing has already been done - is worthwhile, as it would only reduce by some three hours the speed record that Spain has achieved with the present system of processing the provisional vote count. Another uncertain aspect is that of the maintenance and inspection of the thousands of electronic voting devices which would have to be monitored before a new election. The "web" vote is also being studied.

Operations of control of electors and candidates in the electoral process:

In Spain the presence of international observers was not formally requested by any political party during the first elections of the transition thirty years ago. This was because, on the one hand, the now legalized principal political parties already had an infrastructure and organization sufficient to guarantee, with militants and sympathizers, the development of the campaign. They could also ensure proper conduct of the vote by the presence of their legally accredited delegates in almost all of the voting stations in the respective constituencies. On the other hand, the Spanish electoral system introduced from the outset the principle that the sovereignty of the people should translate into active control of the sovereign act of voting by the electors themselves, they being the ones who should preside over and organize the process of voting and vote-counting. The Electoral Law, as subsequently developed, has continued to define these two aspects in greater detail.

Regarding the ability to control candidates in the electoral process, Spanish legislation offers nothing new. The candidacies are submitted to the Electoral Councils in each constituency, which use objective criteria in assessing their validity; it is compulsory also for each political grouping competing in the elections to appoint a Representative to the Central Electoral Council within a maximum of nine days after the election is called, to act as legal spokesperson for the candidacies. This does not mean that these Representatives participate in the deliberations of that organization.

In regard to control of the vote by the electorate itself, it may in fact be new for some people that it is compulsory for the three members of the electoral tables (one President and two Members) to be electors registered on the electoral tables' list. The procedure for their appointment endeavours to avoid any manipulation, as a public drawing of lots is held at a plenary meeting held in each Town Hall between 25 and 29 days after the elections are called. At this meeting three titular members are appointed and six alternate members, who must all be present at the table on polling day in case the titular members are absent. The Presidents are those with the highest educational level. Electors appointed by lot have a legal obligation to take up these duties and the appointments are notified to them by official letter from the Electoral Council, delivered to their home address. Together with this notification they receive an Instructions Manual supervised by the Central Electoral Council. These manuals, like the rest of the electoral documents, are published by the Home Office in bilingual editions for those constituencies, which have Catalan, Basque, or Galician as their official language, as well as in Castilian, which is the official language for the whole State. In Spain the Spanish language does not exist officially in the Constitution, which only refers to what is historically called the Castilian language; it is known abroad as the Spanish language, since in Article Three of the Constitution, the other languages cited are considered Spanish languages.

The appointment of the members of the tables by lot from among the electorate and the guarantees given for them to exercise their mandate in sovereign fashion, which is generally done with a great sense of responsibility, have also been a key factor in ensuring that fraud is actually very difficult at the voting stage, the counting stage, (a public count on the same premises as the vote), and the subsequent delivery by the President and another member of the table, of the original and a copy in a sealed envelope to the nearest Judge. The representatives of the candidacies (who each have a copy of the electoral list of the table and must sign the Minutes of proceedings, of which they receive a copy) also play a part. On public display at the voting station is a notice with the results, whilst a further copy in a sealed envelope is handed by the waiting third member of the table to the Postal worker who collects it at each voting station for onward transmission to the Area or Provincial Electoral Council which will carry out the final count.

In view of the key function of electoral tables, the Electoral Law has introduced specific requirements regarding the number of electors at the tables, and has established sections or areas comprising a minimum of 500 electors and a maximum of 2,000 who are subdivided into tables (in practice the maximum limit of electors per table is set at 1,000 electors). Also a territorial criterion is imposed, that in each municipality, however few inhabitants it may have, at least one table is established even though the electors may be fewer than 500. The electoral list of each table is in alphabetical order of surnames. The Electoral Law also determines that the duration of voting is 11 hours (ample time as voting formalities are carried out very quickly) and voting stations may not close between 9.00 hours and 20.00 hours, even if all the electors on the list have voted.

Operations of financial management of the electoral process: this is an important area of management throughout the phases of the electoral process. According to the Spanish Electoral Law, the State must subsidize not only the actual organizational and logistical expenses incurred by any election but also, in compliance with the Constitutional Law on the Funding of Political Parties, it must subsidize the electoral expenses of the political groupings which compete in elections to the Central, Local, or European Parliaments. The expense of elections to the regional autonomous Parliaments must be borne by each Autonomous Community along the same principles. The Home Office department in charge of running the elections is responsible both for the preparation of the budget and for the administration of public electoral funds according to the precise provisions of the Electoral Law. In fact, the Home Office acts as the administrative body empowered to contract for the external materials and services needed to organize the elections; it also acts simply as the intermediary between the Treasury and the active participants competing in the elections.

The Electoral Law provides - within the financial limit approved by Parliament - that after the 29 days for publication of the decree calling an election, the parties or political groupings which have previously obtained representation, may have an advance of funds equivalent to 30 percent of the total subsidies which they obtained in the previous election. The total subsidy that they receive at the end of the new electoral process, provided they gain representation, will depend on the number of votes. The money is paid by the State after expiry of the period allowed for contentious claims and after submitting (100 days after the ballot) detailed documents covering all electoral income and expenses on accounts opened for this purpose by the candidacies, which can be inspected throughout the electoral process by the Electoral Councils and the National Audit Office. The Electoral Law prohibits private or company donations to the electoral campaign of a political party or grouping of electors other than in a very small individual amount (approximately US $7,000). Should they not obtain representation or obtain fewer votes than those used for the initial computation, candidatures must return the subsidies advanced either wholly or in part.

The expenses subsidized by the State that are connected with the electoral activity of the candidatures are general: voting papers and envelopes; advertising and publicity expenses to gain votes; hire of premises and offices for the campaign; financial indemnities paid to non-permanent staff of the parties, taken on during the campaign; transport and travel expenses of the candidates, leaders and support staff of the campaign; and correspondence or mailing expenses. In addition, bank interest on financial loans legally made to the campaign organization up to the date on which the State pays the total subsidy relating to each candidacy according to its electoral results.

We would clarify that there is a possible modification of electoral legislation currently under discussion regarding the subsidizing of political parties and electoral campaigns, due to problems of corruption. Spain, like many other countries in Western Europe, has suffered from this in recent years. It is generated above all by the secret and illegal financing of some electoral campaigns, where the costs are too high and where the legal limits of maximum expenses established for public funds are not normally observed. The debate is tending towards greater possibility of financing by private persons or companies, by substantially raising the maximum level of such contributions currently contemplated in the Law. However, should the electoral legislation be modified, private financial contributions would always be subject, by parliamentary consensus, to compulsory publication and control.

The Transition from Territorial State

It would not be possible to conclude a description of the Spanish political regime and of its electoral system without briefly mentioning something just as important historically as the transition from authoritarian state par excellence to an exemplary democratic one. We refer to the transition that was made in parallel with that previously described, from super-centralized State to strongly decentralized State - a transition which at the political and sociological level has undoubtedly been more difficult and complex than that from dictatorship to democracy. Perhaps it was the toughest problem that the 1978 Constitution had to tackle.

In practice, for many Spanish people whom we could call ultra-conservative, just as for many who have a progressive political ideology, it has been, and still is, difficult to accept the recognition accorded to nationalists, linguistic diversity, and the creation of self-government in the various Spanish regions which have been invested with ever wider powers. Furthermore, the arrival of democracy facilitated an explosion of strong nationalist claims, which have since been largely assimilated into the political system through the electoral mechanisms underpinning the decentralization process. Notably, the system even allows HB, the political candidacy linked with the terrorist group ETA, to stand lawfully in all elections (legislative, municipal, Basque Country/Autonomous, and European). Everyone has always accepted the results, although for some years they have shown a strong decline in the HB vote.

Within the framework of all these tensions, the Spanish pattern of decentralization was original, in that it took account of the existence of different demands for autonomy vis-a-vis central power. These were very strong in Catalonia and in the Basque country, moderate in other areas, and practically nil in quite a number of regions, and it was able to devise a long-term strategy for the steady harmonization of the self-governing capacities of each territory.

On these principles the division of Spain was organized into 17 Autonomous Communities, each having a Parliament elected by universal suffrage. These parliaments were instituted successively, the transition from the State model beginning with the autonomous elections of the Basque Country and of Catalonia held at the beginning of 1980. The parliaments of Galicia and Andalucia were elected subsequently, in 1981 and 1982 respectively. These four regions are the ones called "historic nationalities". The remaining 13 autonomous parliaments were established as from the autonomous elections held simultaneously with the local elections of 1983.

In the State decentralizing process, one of the basic features of the general harmonization was precisely the adoption of a similar electoral system throughout the land: the election of the autonomous parliaments (although their deputies vary in number) is governed by the same rules for distribution of seats as for the election of deputies to the central Parliament (las Cortes). Each Autonomous Community has its own Electoral Law, which should align with the General Electoral Law in its basic procedures.

Based on these autonomous parliamentary assemblies elected by universal suffrage, a phenomenon of moderately developing regionalism has been generated in many places where this consciousness did not exist. However, there is also a phenomenon of integration of nationalisms within a robust and organized framework, which allows the central and autonomous powers to negotiate on bases of legitimacy, which both obtain at the ballot box. It is these negotiations which have allowed a high degree of decentralization (much greater than that within many states formally termed "federal") which today is accepted by the majority.

Finland: Candidate Choice and Party Proportionality

The Finnish electoral system was introduced in 1906. Elections were held the following year, which were the first free proportional elections for both men and women. In 1917 Finland became independent from Russia, and the founding Constitution of the new Republic was put into force in 1919; later a variant of parliamentarism known as semi-presidentialism was developed. Since 1906, all women and men have been eligible to vote and to be nominated in elections. The age of eligibility has been successively lowered from 24 in 1906 to 21 in 1944, to 20 in 1969 and to 18 in 1972. One distinctive feature of Finnish elections is the exceptionally high numbers of ballots cast in absentia by post. At the 1995 election, 43.4 percent of the valid votes were so cast.

The Finnish parliament consists of 200 MPs elected from 15 districts. In all districts, except on the Swedish-speaking Åland Islands, the allocation of seats to parties (including electoral alliances) is proportional to the votes following a d'Hondt system of party list Proportional Representation, see Mixed Member Proportional. Before 1954, voters had to choose between candidate lists (a list included a maximum of two candidates and one alternate); later changes to the system mean that it is now possible to vote for one individual candidate only. This change transformed the Finnish electoral system into a rare type of list system, which obliges voting for individual candidates.

The election of candidates from the party list is not predetermined, but depends entirely on the number of individual votes cast for each candidate. The voter picks the allotted number of his or her candidate (the list of candidates, each with an identifying number, hangs in front of the voter) and writes it down on the ballot. As a result, the election is not exclusively a competition between parties; it is also a competition between single candidates on the party list. Neither is the electorate given the option to vote for a party per se; but only for individual candidates nominated, but not ranked by a party or a non-party list.

While the Åland Islands district elects a single member, the other 14 districts are all multi-member. The district magnitude is determined by the population size, which favours the constituencies in the rural north and east. Proportionality is still high in overall parliamentary results, although variation between constituencies in this respect is large. In general, urbanised constituencies are more proportional, and more rural areas produce more disproportional results.

As the d'Hondt formula of allocating seats favours large parties, in Finland small parties usually take the opportunity of joining an electoral alliance with one or more parties. Electoral alliances are made at the district level, which means that one party can join different alliances in all 14 districts; the alliances therefore have varying degrees of success. In addition, according to the electoral law of 1969, a candidate can only be nominated in one constituency. Before that a candidate could be nominated in all districts, the optimal electoral strategy for a charismatic small-party leader. Most small parties join electoral alliances, and without this option proportionality between votes and seats would, to some degree, be weakened.

Indonesia: Continuity, Deals and Consensus

The development of political institutions able to provide stable and effective government has been a daunting challenge in Indonesia, a huge and ethnically diverse country of nearly 20,000 islands whose unity was based on common resistance to colonialism. Political identity in Indonesia is a complex subject, in the past often based on links to different strands of Islam, to a more secular nationalism, or in some areas to Christianity—to which assessment of leadership qualities and the impact of corruption now, in 2004, appear to have been added. Devising electoral systems that are inclusive and effective in the context of the Indonesian unitary state has never been easy.

The first general election in Indonesia after the 1945 proclamation of independence took place in 1955. A PR system using 15 regions was adopted without challenge. Seats were distributed in proportion to population, with a small extra allocation for Outer Island regions. The Largest Remainder Method using the Hare Quota was adopted. Parties or organizations could nominate lists, and individual candidates could also be nominated. Voters could vote either for a list or by writing in the name of one candidate.

The resulting legislature included representatives of 27 parties and lists, plus one individual member. The four largest parties all received between 16 per cent and 23 per cent of the vote. Not only was no single party able to command a majority in the legislature; not even two parties were. It was difficult to form governments, and their ability to retain the confidence of the legislature was limited. The Constituent Assembly, elected shortly afterwards to draw up a permanent constitution, had a similar political balance and failed to reach agreement.

General loss of confidence in political institutions and rebellions against the unitary state led President Soekarno to impose an authoritarian regime in 1959. This lasted until it was replaced by the New Order of President Soeharto in the mid-1960s, which established virtually complete dominance of the executive, legislative and judicial branches. Elections took place, but campaigns were heavily restricted, many candidates were disqualified, and the rules were applied disproportionately against opponents of the government. The desire for complete central control over the choice of candidates contributed to the choice of closed-list PR. Soeharto sought to allay fears of Javanese political domination, and Java received only just over half of the seats to be elected, despite having over 70 per cent of the registered electorate in 1955. Although this figure has fallen, it was still 61 per cent in 2004.

The Transition to Democracy: The 1999 Elections

After the Soeharto regime fell in 1998, new electoral legislation was finalized in late January 1999. The electoral system—described as a ‘proportional system with district characteristics’—was unique. It was the clear product of incremental political negotiation against a time deadline. This agreement was reached in the legislature by the parties of the Soeharto era, which were under pressure from the new parties and others outside the negotiations, in addition to defending their own positions and coming under pressure from their own power bases. Given these pressures, it is unlikely that the end result of the negotiation could have been substantially different.

At the June 1999 election, each voter cast a single vote for a political party. The 27 provinces were retained as electoral districts, ranging in magnitude from four to 82 seats. The number of seats won by each party in each province was determined using the principles of PR, and each candidate on each party list was linked by the party to one of the second-tier districts (kota, city authorities, and kabupaten, authorities in non-city areas) within the province. Because of deep-rooted concern to maintain the unity of the state, there were provisions that made it impossible to register a specifically regional party. The law on political parties required all parties contesting the elections to be organized in at least nine provinces.

The legislation was not clear on essential detail. The method for converting votes cast into seats gained was not included. Nor were there rules for identifying which candidates from a party’s list would occupy the seats gained by that party. These issues were resolved only at a very late stage.

The final seat allocation regulations retained the Largest Remainder Method using the Hare Quota. The allocation of candidates to seats won was much more complex. In practice, few central party leaderships even complied with the regulations, and leaderships took de facto powers simply to tell the Election Commission which candidates had been elected to the seats their party had won.

The 1999 elections were nonetheless judged to have been the first since 1955 to be acceptable overall, despite specific or localized concerns. Five parties gained more than 2 per cent of the vote: their relative strengths varied widely in the different parts of Indonesia. Sixteen other parties gained representation.

Reflection on the 1999 elections was rapidly overtaken by a full review of the 1945 constitution. The completion of this review in 2002 led to fundamental changes, including the introduction of the separation of powers, the principle of checks and balances, direct election of the president and vice-president, and the establishment of a regionally based elected second chamber with very limited powers. Four portmanteau amendments were passed to the constitution, completely changing the way in which the institutions will work in future, and five new laws—on elections, presidential elections, political parties, the structure of elected bodies, and the establishment of the Constitutional Court—were passed. Indonesia now falls recognizably within the mainstream of the family of presidential democracies.

Single-Member Plurality Fails to Find Favour for Elections to the Legislature

After 1999, there was considerable advocacy of a single-member district (SMD) plurality system among the media and in academia in particular, as the accountability of elected members was widely perceived to be lacking in the legislature elected in 1999. Even if the 1999 electoral system is viewed not as a political deal but as a brave attempt to marry the principles of List PR and the accountability of elected members to the electoral district, no constituency–member link was created in practice.

However, simulations made after the 1999 elections suggested that a plurality SMD system would be likely to produce results in Indonesia that were more disproportional than almost anywhere else in the world. Steadily worsening relations between the legislature elected in 1999 and many academic, media and civil society actors also meant that support for SMD systems by the latter became steadily less persuasive. It became evident that a plurality system would almost certainly fail to reflect the diversity of Indonesia, that introducing an acceptable districting process for the 2004 election would take time and involve considerable difficulty, and that plurality systems were not likely to favour the election of women.

The 2004 Electoral System

The new constitutional requirements agreed in 2002 state that the participants in elections to the lower chamber of the legislature (the National Assembly) are political parties, thus limiting the available options for the electoral system in the new electoral law. The government’s draft election law provided for a PR system using multi-member districts, responding to the pressure for greater accountability by proposing open lists and the dividing up of larger provinces. This basic form was finally adopted, with multi-member districts of a magnitude of between three and 12 seats to be drawn up by the Election Commission. Subsequent debate led to multi-member districts whose magnitude is towards the higher end of this range. The restricted open-list system finally agreed requires voters to vote for one party and, if they wish, one candidate from that party. However, this will only result in the election of a particular candidate out of the order in which names appear on the party list if that candidate gains more than a full Hare Quota of individual votes—which made its likely effect minimal, as proved to be the case in practice in the 2004 elections to the legislature.

As a result of the creation of the regional chamber, some parties argued the case for ‘one person, one vote, one value’ (OPOVOV) for the legislature, with the same population for every seat, while others backed the retention of a representational bias in favour of the Outer Islands. The final compromise is a complex formula basing the number of seats for each province on a minimum of 325,000 population per seat in small provinces and a maximum of 425,000 population per seat in large ones, with a minimum of three seats per province.

The central party leaderships showed little inclination to relax their hold on their parties. The larger parties toughened the requirements for parties to participate in both the 2004 and subsequent elections.

The effective prohibition of regional parties has been strengthened. The issues of open or closed list, OPOVOV, the balance between Java and the Outer Islands, and party participation were all negotiated between the parties when the final deal was struck. A broadly-based campaign did, however, lead to the adoption of a ‘maybe-quota’ for gender representation: parties are required to ‘bear in their hearts’ the desirability of including at least 30 per cent women candidates on their lists. While there is no enforcement provision, this proved an important tool to encourage more women candidates, and 12 per cent of the members of the 2004 legislature are women-a significant improvement on 1999.

The 2004 legislative election results reflected both change and continuity. The same five parties that polled more than 3 per cent in 1999 did so again, and were joined by two more. Seventeen parties were represented altogether.

Elections to the Regional Chamber: SNTV Springs a Surprise

The constitution provides that candidates for the new regional chamber (the Regional Representatives’ Council) should be individuals, not parties. Four members are to be elected per province. The draft law proposed the Block Vote system, clearly designed to advantage parties with support outside Java where provinces are smaller. SNTV was proposed as an alternative by the party that was strongest in Java, and was agreed as part of the final deal.

The first election to the regional chamber took place in 2004 and demonstrated a known weakness of SNTV: with an average of 30 candidates contesting the four seats in each province, many candidates were elected with less than 10 per cent of the vote. However, strong campaigns by women candidates meant that an unexpected 21 per cent of the members of the new chamber are women—a level unprecedented in a freely elected body in Indonesia.

Direct Presidential Elections

The president and vice-president are now directly elected, with candidates pairing up to form tickets. A two-round majority run-off system is used, with the aim of ensuring that the successful candidates have sufficient support across a large and diverse country. For a ticket to be elected on the first round, it must not only poll an absolute majority of votes cast but also meet a distribution requirement of 20 per cent of the vote in at least half the provinces. While a majority winner will almost certainly achieve this, the requirement prevents a ticket whose support is solid in Java and minimal elsewhere from winning an election in the first round. In the first direct presidential election in 2004, five tickets contested the first round in July, with none polling over 35 per cent; in the second round in September, Susilo Bambang Yudhoyono gained victory with 61 per cent of the votes.

The Political Reality: Negotiating a Deal

The 1999 electoral system agreement had to be acceptable both to the parties of the New Order, which still held the levers of power, and to the new parties outside on the street. The constitutional review that followed the 1999 election also required agreement across the political spectrum. The 2004 election law is yet another deal, similar in principle to that of 1999, but with important differences of detail. Each time, there were a limited number of practical solutions given the inherited traditions and the political background and positions of the actors. However, there are positive signs for democracy in the new Indonesian institutional framework: it is fortunate that some long-term vision existed alongside the inevitable perceptions of short-term political advantage among the parties and individuals who shaped the changes.

Republic of Ireland: The Single Transferable Vote in Action

The Irish lower house of Parliament, Dáil Éireann, is elected by the STV system—proportional representation by means of the Single Transferable Vote. This relatively unusual system owes its origins to the circumstances of the Republic of Ireland’s achievement of independence in 1922. The departing rulers, the British, wanted some form of PR in order to protect the Protestant minority, while the new state’s political elite favoured PR in principle. With neither having much awareness of PR list systems, STV was adopted by agreement as the electoral system and has remained the electoral system ever since.

The Dáil is of central importance in the Irish political system. It elects the government, which needs to retain majority support in the Dáil in order to survive. Much less important is the presidency, although, unusually for a parliamentary system, the president is directly elected. Elections for the presidency take place under the Alternative Vote (AV) system.

The 166 members of the Dáil are elected from around 40 constituencies, each returning three, four or five members. Voting is straightforward: voters merely indicate their favoured candidate (by writing ‘1’ beside that candidate’s name on the ballot paper), and can go on to indicate their second and third choices and so on in the same way. Voters can rank candidates not only within but also across parties. Although most vote along party lines, it is not necessary to do so, and some vote along geographical lines, that is, they give their highest preferences, regardless of party, to the candidates from their own local area. The counting process, especially the distribution of ‘surplus’ votes, looks complicated to the uninitiated, but it is worth emphasizing that the voters do not have to be familiar with all the details; they need only to know how to cast their vote and to be satisfied that the counting process is ‘fair’ and transparent.

The electoral system is entrenched in the constitution and consequently cannot be changed without a referendum. On two occasions (1959 and 1968) the largest party, Fianna Fáil, instigated a referendum to replace STV by the British FPTP system, using the argument each time that any kind of PR was likely to create a problem of unstable coalition government. The proposed change was rejected by the voters on each occasion, by margins of 52 per cent to 48 per cent in 1959, and 61 per cent to 39 per cent in 1968.

On the basis of the criterion of stable government, anyone evaluating the record of STV in the Republic of Ireland would not, in fact, see its performance as a problem. Since the mid-1940s, governments (both coalition and single-party) have lasted three, four or five years, the only exception being a short-lived period of instability in the early 1980s. The voters, through their ranking of candidates of different parties, are able to indicate their wishes regarding potential coalition partners for their preferred party.

STV has generally delivered highly proportional outcomes, with Fianna Fáil receiving only a modest ‘bonus’ (around 48 per cent of the seats for 45 per cent of the votes at elections over the period 1945–92). However, the small size of the electoral districts (four seats per constituency on average) creates the potential for the largest party to reap a benefit if it can attract second- and third-preference votes from supporters of other parties. In 2002 the least proportional results ever, Fianna Fáil won 41 per cent of the votes and 49 per cent of the seats.

The system continues to allow representation to small parties and to independents, 13 of whom were elected in 2002. While many PR systems enable small parties to win seats in the Parliament, STV seems to give an unusual opportunity to independent candidates to do the same because of its essentially candidate-centred rather than party-centred nature.

Much of the praise and criticism of STV in the Republic of Ireland hinges on the same factor, namely the power it gives to voters to choose among candidates of the same party. This creates intense intra-party competition, especially among candidates of Fianna Fáil, which nominates between two and four candidates in each constituency. Statistics show that more incumbent Fianna Fáil MPs lose their seat to a running-mate than to a candidate of another party.

Critics argue that, as a result, incumbents become over-active at constituency level in order to curry favour with the voters and do not spend enough time on politics at national level, for example, on scrutinizing the government or discussing legislation in committees. They argue that this has an adverse effect on the calibre of Irish parliamentarians (in that individuals who could make a contribution at national level are discouraged by the likely casework load they would have to discharge if elected) and that it leads to short-termism and undue regard for localism in government thinking. They suggest that internal party competition for votes may lead to divided, incohesive political parties.

The defenders of the system, in contrast, see voters’ opportunity to choose among candidates of their party as a virtue. They argue that it allows the voters to replace incumbents by more able and more active newcomers and that, at a time of decreasing interest in conventional politics, this gives MPs a strong incentive to keep in close contact with the voters and thus fulfil the role of linking citizens to the political system. They maintain that there is no evidence that Irish MPs are of lower calibre than those elsewhere and that the Republic of Ireland’s recent record of impressive economic growth shows that there cannot be too much wrong with the behaviour of governments. They also point out that the Irish political parties are extremely cohesive and disciplined in their behaviour in Parliament, with no factions or recognizable subgroups.

In 2002 an all-party parliamentary committee considered the arguments for and against changing the system. It concluded that the public was strongly attached to STV, that a change to any other system would reduce the power of the individual voter, and that some of the alleged failings of the political system for which critics blamed STV were caused by other factors. As this conclusion indicates, there is no significant body of opinion in favour of amending or replacing the existing system.

Any evaluation of STV in the Republic of Ireland needs to take account of the characteristics of the country. It is a small country in terms of both area and population, and the ratio of MPs to population (about 1 : 20,000) is relatively high by international standards. This may foster closer links between MPs and their constituents, regardless of the electoral system, than are likely in a larger country. In addition, the Republic of Ireland is a prosperous, highly educated society where the political system as a whole is well established and is universally regarded as legitimate. Irish society does not have any significant cleavages (for example, ethnic, linguistic or religious).

For all these reasons we need to be careful about drawing firm conclusions about how STV would operate in other contexts. We can, though, say that there is no sign that the electorate in the Republic of Ireland would like to replace it by any other system.

Israel: Electoral Reform in Israel

On March 18, 1992, on its last day before disbanding, the Israeli parliament, the Knesset, changed the electoral law in Israel. This momentous institutional change was implemented in the fourteenth general election on May 29, 1996.

The initiative to reform the electoral law emanated from widespread dissatisfaction over government performance. A grass-roots movement led by prominent law professors and Knesset members attributed the stalemate in Israeli politics during the 1980s primarily to coalition politics. Small parties, particularly the religious ones, gained disproportionate influence in the coalition formation process, thus weakening the discretionary authority of the Prime Minister over the formation of national public policies, and bestowing them with larger than deserved shares of public resources and symbolic commitments. The institutional change was supposed to remedy this situation.

In a recent paper, Nachmias and Sened (1998) show that the institutional reform in the electoral law significantly decreased the electoral strength of the big parties and inevitably augmented the bargaining power of the religious and other small parties. From the reformers' perspective, the outcome of the electoral change was counter-intuitive. However, from a theoretical perspective, the consequences were to be expected. In the first section, I examine the electoral rule used in Israel between 1951-1992, and discuss the political reasons for the widespread dissatisfaction with this institutional arrangement. In the second section, the major attributes of the new law are described along with the unrealizable expectations that it would constitute a significant improvement over the previous law. The problems inherent in the new law are discussed from a conceptual perspective in the third section. In the last section, the general implications of the reform are addressed in the context of the search for pluralistic, democratic institutional designs.

The Electoral System in Israel Before 1996

The electoral law practiced in Israel from 1951 to 1992 was one of the purest forms of proportional rule. The entire Israeli electorate was treated as a single district. The number of seats that each party in the Knesset gained was almost exactly proportional to the number of votes the party obtained in the general elections. The minimum number of votes needed to enter the parliament was one percent of the votes (since 1992, 1.5 percent), a very low threshold of entry compared to similar electoral systems.

One notable effect of this pure form of proportional electoral rule was that the Israeli Knesset was always composed of a multitude of parties. In the thirteen elections held between 1949-1992, no party ever obtained a majority of the seats in the Knesset. This required the largest party in the Knesset to enter a bargaining process of coalition formation after every electoral campaign in order to form a new government.

Coalition formation in multi-party systems involves a tedious bargaining process over two types of payoffs: office-related side payments and policy agreements (Laver and Schofield 1990; Sened 1996). In multi-party systems, the coalition forming party must strike a balance between office-related side payments and policy-related payoffs. Any government can pursue only one policy that rarely satisfies all the partners in the ruling coalition. The coalition forming party must win the support of its smaller partners. Side-payment are allocated to them to compensate their dissatisfaction with policies that the government as a whole decides to pursue. This dissatisfaction often leads to the breakdown of coalition governments, when the office-related side payments no longer compensate for the policy compromises of the different coalition partners (Mershon 1996; Sened 1996).

The New Electoral Law

Towards the end of the 1980s a grass-roots movement, the Public Committee for a Constitution for Israel was formed to advocate political and electoral reforms. The group's leadership submitted a detailed proposal for reforms, including specific recommendations pertaining to issues of individual rights; a formalized structure of checks and balances between the legislative and the executive branches of government, and a new electoral law. Despite of the publicized objection of most political scientists in Israel and a sizable number of legislators representing different parties, the Knesset, after considerable political pondering, maneuvering and delay tactics modified the original movement's proposal and changed the electoral law. The success of the movement has been attributed to its outstanding public campaign in terms of scope, resources, and visibility, reinforced by the endorsement of leaders of the two major parties.

The new electoral law includes two major provisions that are supposed to strengthen the Prime Minister in the process of forming a coalition following a general election. First and foremost, the Prime Minister is elected directly by the eligible voters. Voters enter two ballots in the poll. On the first ballot, they vote for the party of their choice and on the second they vote for their most preferred candidate for Prime Minister. Under the old electoral law, following the election, the President of Israel called all the newly elected Knesset members and consulted with them before asking one of them to try to form a coalition. In practice (with a single exception), the president asked the head of the largest party to attempt forming a coalition government. Under the new law, the Prime Ministerial candidate receiving more then fifty percent of the votes is popularly and directly elected. If no candidate receives more than fifty percent of the votes, a second round is held. Blank and invalid ballots are not counted.

Consequently, one of the two contenders is assured of obtaining more then fifty percent of the votes. Under this procedure, the Prime Minister may not be the head of the largest party in the Knesset. In fact, in the 1996 election Netanyahu won the election for Prime Minister while his Likud party gained 32 seats in the Knesset. Two seats less than Labor, which gained 34 seats while its head and Netanyahu's contender for the Prime Ministerial election, Shimon Peres, lost in the election for Prime Minister.

To further strengthen the power of the Prime Minister and ensure the stability of coalition governments, the new law diminished considerably the potency of the long-standing parliamentary institution of vote of no confidence. Under the new law an absolute majority of Knesset members (61 members) is required to approve a vote of no confidence, compared to the old rule where a simple majority of the members present in the plenary was sufficient to pass such a vote. Most significantly, however, under the new law, if an absolute majority supports a vote of no confidence, not only is the government ruled out of power, but the Knesset is dissolved as well. This change constitutes a strong disincentive to legislators to support a vote of no confidence. To bring down a coalition government without dissolving the Knesset, a vote of no confidence must be supported by at least 80 Knesset members.

Legislated Rule was not to serve the Purpose

There are three reasons to expect that the institutional change will fail the major purpose for which it was legislated. The first is well- known and has been pointed out by scholars and a few prominent elected officials prior to the change. The new electoral law enables small parties to pressure the big parties to accommodate their policy preferences in three rounds: preceding the first-round of the election, again before the second-round of the election, and still again during the bargaining process for the formation of the coalition government.

Under the old law, the small parties could pressure the big ones only during the coalition formation process, and solely if they were genuinely pivotal. Under the new rule, the small parties in general and the religious parties in particular, are virtually guaranteed the pivotal status in the second round. Preclusion of a second round would make them pivotal in the first round. This institutional change increases considerably the likelihood of small parties, in particular the religious, to become genuinely pivotal thus augmenting their bargaining power.

The second reason to expect the new law to defeat its purpose is directly related to another well-anticipated outcome: big parties are bound to lose Knessset seats to small parties. Since the Labor Party lost its dominance status, the Israeli party system has turned into a bi-polar system in which two major parties compete, with the support of their satellite parties, over the control of parliament and hence the coalition government. Typically, the head of the largest party was given the first opportunity to form a governing coalition. The voters, fully aware of this institutional practice often voted strategically to increase the chances of the head of the big party of the Knesset block they preferred to get the first opportunity to form a coalition. Voters could be closer in their ideological persuasion or policy preferences to one of the small parties in the Knesset-block and yet cast their ballot to the biggest party in the block to increase its chances to get the first opportunity to form a governing coalition.

The new electoral law eliminated the incentive to vote for the biggest party in the block. Under the new law, voters can cast a ballot for the head of the party that leads the parliamentary block they prefer, and then vote sincerely for the party of their choice. Inevitably this leads to added fragmentation in parliament inasmuch as it steers voters to cast their vote to small parties instead of one of the biggest parties. This added fragmentation intensifies the governability problems inherent in coalition governments in different ways:

First, as Schofield (1995) demonstrates, for all practical purposes, a necessary condition for a non-empty core in two dimensional policy spaces, like the Israeli policy space, is that one central, dominant party must have a considerable advantage in size and occupy a central position in parliament.The diminished electoral size of the big parties, resulting from the change in the electoral law, and the inevitable increase of the power of small and medium-size parties, virtually eliminates the possibility of a stable core in the Knesset's policy space. A dominant, core party can pay considerably less office-related side-payments to its coalition partners than less central smaller parties who may attempt to form a coalition. The advantage that the core, dominant party has in the coalition formation process empowers it to pursue relatively consistent, long-term policies and reward coalition partners with secondary portfolios in order to obtain their support of the government and its policies. The low likelihood for a dominant party to emerge under the new electoral law impairs the ability of the government to maintain consistent policies. Concurrently, it raises the price that coalition-forming parties have to pay to secure the support of their partners in government.
The other reason that the loss of Knesset seats by the two block leaders to smaller parties is likely to reduce the share of portfolios held by the coalition forming parties has to do with straightforward arithmetic. The coalition forming party must obtain the support of at least 61 Knesset members in order to present the coalition to the vote of investiture, a central institution in multi-party parliamentary systems and an indispensable proviso for a coalition to become a formal government. The bargaining unit in multi-party parliamentary systems is the party. Each party joining the coalition presents, at the coalition formation, bargaining process, its policy demands as well as its office-related preferences. Given that the government can pursue only one policy position, the coalition forming party must compensate coalition partners with portfolios to the extent that they compromise their policy preferences. This implies that the number of the coalition partners should be positively correlated with the cost of the coalition formation in terms of the portfolios that the coalition forming party must give away to its partners. The reduced number of Knesset seats that potential coalitions forming parties are expected to have due to the new law, necessitates an increase in the number of coalition partners in order to form a minimum winning coalition and pass the vote of investiture in the Knesset. Thus the new rule is expected to raise the number of portfolios to be allocated by formers of coalitions to their partners.
The third reason to expect the new law to defeat its purpose is the remarkable erosion it introduces in the force of the parliamentary institution of the vote of no confidence. From an effective governance perspective, the importance of the vote of no confidence as a parliamentary practice has been well-explained by Huber (1996: 279): "By allowing the Prime Minister to make the final policy proposal, confidence vote procedures give the Prime Minister substantial influence over final outcomes, even when these procedures are not invoked." In other words, by invoking the vote of the no confidence procedure, the Prime Minister can discipline coalition partners to vote with the government even if they disapprove of a particular policy in question. Since under the new law Knesset members lack the incentive to pass any vote of no confidence, the Prime Minister has lost an important governing resource. Furthermore, the vote of no confidence ceased being a credible threat that excessive budget and policy demands by small coalition partners may lead to the downfall of the governing coalition. This, in turn, leads the smaller parties to raise their demands for policy-related payoffs from coalitions forming parties.

In sum, there are three theoretically grounded reasons to expect the institutional change to accomplish the opposite of what it was intended to do. Instead of reducing the fragmentation in government and decreasing the bargaining power of small parties, the new electoral law is expected to increase fragmentation as well as the power of small parties, particularly the religious parties. First, the new law institutionalizes more opportunities for small parties to bargain with Prime Ministerial candidates, and then with the Prime Minister-elect. To raise their chances of wining the election, candidates will promise a variety of payoffs to small parties in return for their support in the first and the second round and during the coalition formation process. Second, the new rule reverses the incentive of voters to vote strategically for the big party in the Knesset block. Inevitably the big parties would lose Knesset seats, considerably reducing the likelihood that a major, central party would capture the parliamentary core and pursue relatively consistent policies. Moreover, the coalition forming parties would have to rely on more partners to form and maintain governing coalitions. This would increase both fragmentation in government and coalition payoffs to small parties. Finally, by jeopardizing the role of the confidence vote procedure the new law would impair both the governance capability of coalitions and the effectiveness of parliamentary oppositions.

India - First Past the Post on a Grand Scale

India remains by far the largest democracy in the world, with over 670 million electors in the parliamentary election of 2004. Its parliamentary government and FPTP electoral system are a legacy of British colonialism, which ended in 1947.

The British introduced self-government to India in stages, and it was not until the end of colonial rule and the adoption of the Indian Constitution in November 1949 by a Constituent Assembly that universal suffrage was achieved.

The Constituent Assembly, which comprised eminent jurists, lawyers, constitutional experts and political thinkers, and laboured for almost three years, debated at great length which electoral system would be best suited to India before finally choosing to retain the FPTP electoral system. Various systems of proportional representation were considered and attracted many advocates, given India’s extremely diverse and multi-ethnic society, but FPTP was chosen, mainly to avoid fragmented legislatures and to help the formation of stable governments—stability being a major consideration in a country emerging from immediate post-colonial communal bloodshed and with widespread poverty and illiteracy.

Under the Indian Constitution, voters elect a 543-member Lok Sabha, or lower house, from single-member districts. By contrast, the upper house of Parliament, the Rajya Sabha or Council of States, and the corresponding upper houses of some states, are indirectly elected by members of the state legislative assemblies. There are also a president and vice-president who are indirectly elected by the members of parliament and state legislative assemblies.

General elections are held once every five years, but the president may dissolve the Lok Sabha on the advice of the prime minister before its term is over, as in the recent case of 2004, or if he or she is convinced that no stable government can be formed, as in 1991. The prime minister holds office for as long as he or she can command a majority in the Lok Sabha. All the successive Congress Party governments which ruled India continuously until 1977 served for almost five years, close to the maximum allowed in the constitution. From 1977 to 1997, governments were less stable, and a number of prime ministers had to resign as a result of party splits or votes of no confidence before completing their full term. Since 1997, a period of stability seems to be emerging again, now under coalitions of parties.

All these political environments have arisen from the same FPTP electoral system. The major effect of the electoral system until 1977 was to guarantee majority governments based on a minority of voter support. The FPTP electoral system initially resulted in the ruling Congress Party securing stable majorities in the Lok Sabha, usually against a fragmented opposition. This fragmentation was characterized by a rise in popularity for regional and state parties in some areas. When the opposition parties combined to form coalitions and started putting up common candidates against the Congress candidates (as was the case in the 1977 and 1989 general elections), the Congress majorities vanished. Moreover, the nature of the system meant that small changes in share of the vote often had a dramatic impact upon the number of parliamentary seats won, as the following table, relating votes for the Congress Party to the number of seats won at successive elections, illustrates.

The Congress Party’s Performance in Indian General Elections: The dramatically large effect of the FPTP electoral system on the number of seats with slight changes in voting

The same disproportionality between the share of votes obtained and the share of parliamentary seats won under the Indian FPTP electoral system can be seen in the case of the other major political party, the Bharatiya Janata Party (BJP), which led a coalition government up to 2004.

Thus the overall results of elections to the Lok Sabha have not been anywhere near proportional. Support can often be divided by setting candidates of the same caste, religion or region against each other. In this context, FPTP gives an incentive to electoral participants to encourage multiple candidacies by their opposition, and its effect can be to produce a winner who has much less than an absolute majority of the total vote. However, despite the divided nature of India’s multi-ethnic democracy, the electoral system has retained a considerable degree of support, due in part to the practice of reserving seats for socially underprivileged and historically disadvantaged groups known as scheduled castes and scheduled tribes. These communities are thinly spread all over India, and the classical operation of FPTP would have resulted in them getting a comparatively very small number of parliamentary seats. The constitution, however, reserves districts for them in proportion to their numbers in the population, thus reserving 79 seats for the 15 per cent scheduled castes population and 41 seats for the 8 per cent scheduled tribes population. In these districts, although all electors have voting rights, only a member of the scheduled caste or tribe may stand for election. This has ensured that their parliamentary representation is in line with their proportion of the population.

A constitutional amendment which seeks to reserve 33 per cent of seats for women representatives at the national- and state-level legislatures has long been debated, but without any success so far, although 33 per cent of the seats have been reserved for women at the Panchayat (district) level, the third tier of government, since 1993. The depth of popular support for the integrity of the electoral system became evident in 1977 when the election of the incumbent prime minister, Indira Gandhi, was set aside by a court after Congress had won a two-thirds legislative majority in 1971. She responded by curtailing fundamental constitutional rights for two years (1975–77), an authoritarian interlude in India’s otherwise unbroken history of competitive democracy. In the 1977 elections, her government lost power through a fair poll, signalling the unwillingness of India’s voters to accept undemocratic practices.

For a period of 20 years, from 1977 to 1997, the FPTP electoral system seemed to have ushered in an era of instability, principally because of the formation of coalitions without common principles and the pursuit of narrow self-interest by political parties. The non-Congress opposition parties (without the communists) took over in government in 1977 by uniting into a composite entity, the Janata Party. It split within two years. In December 1989, a successor party, the Janata Dal, came to power, supported by the communist parties and the Hindu revivalist Bharatiya Janata Party (BJP); this government lasted ten months. At the general election of 1996, no party was able to form a stable government. The BJP won 161 seats and the Congress 140. But the strength of the electoral system re-emerged in 1999 when a firm alliance of parties under the leadership of the BJP was able to form a government and almost complete its full term. Similarly, after the May 2004 general election, the Indian National Congress Party, along with left parties and others, formed a coalition government at the national level.

In 2000, the government of India established a National Commission to Review the Working of the Constitution. This commission’s consultation process considered whether various provisions relating to the electoral process in the constitution should be amended or expanded. Its report, submitted to the government in 2002, recommended against any constitutional change in the electoral field, emphasizing that such changes as were needed could be brought about by amendments in the ordinary electoral legislation and even by subordinate legislation or executive instructions. However, the National Commission also observed that, at the last three general elections at national level, an average of two-thirds of Indian MPs had been elected under FPTP without a majority of 50 per cent plus one and with a plurality only, and considered the questions this raises about the legitimacy of representation. As a consequence, and in the context of the nationwide introduction of electronic voting which then took place in 2004, the National Commission recommended that the government and the Election Commission of India conduct a careful and full examination of the introduction of a Two-Round system, with the second round conducted between the two leading candidates in each district on the day after the first round. The report of the Election Commission of India following the 2004 election did not follow up on this proposal, although it did recommend both the introduction of a ‘none of these candidates’ option on ballot paper and the abolition of the provision by which one person is able to stand in two different single-member districts.

The FPTP electoral system is often said to work best in countries where there are two major political parties. In India, by contrast, the Congress Party ruled continuously at the centre from 1952 to 1977 without any viable opposition. This monopoly ended in 1977. From single-party dominance, the pattern on the political arena changed, first to one of a competition between a single party and a coalition of parties, and from there to a competition between two coalitions of political parties—a trend that continued at the 2004 general election. The BJP started its upward mobility in the Indian Parliament with a shrill Hindu agenda, but after one full term in office the imperatives of electoral politics compelled it to scale down its ultra-rightist militant stance. It had to adopt an inclusive agenda, enabling it to appeal to Muslim, tribal, backward class and other Dalit (downtrodden) voters—who were once considered to be in the exclusive domain of the Congress Party.

Jordan - Electoral System Design in the Arab World

The electoral system issue has become the focus of one of the most heated and controversial debates in Jordan since multi-party politics was re-introduced by King Hussein. The November 1989 general election was conducted in an environment where political parties were banned, as had been the case since the early 1960s, but Muslim Brotherhood and pro-monarchist independents were easy to identify. For these elections, the first competitive ones for nearly thirty years, Jordan used the Block Vote electoral system, see Block Vote, which the British had utilised in the territory in the immediate post-war period, to elect their 80-member legislature. Out of these seats, eight were reserved for Christians and another three for Circassians, see Representation of Minorities.

The country was divided into 20 constituencies, returning from two to nine MPs each, but the disparity in size between constituencies returning the same number of MPs was considerable. For example, both the Fifth District of Al-Assima and the constituency of Maan returned five members to the House of Deputies, but the Al-Assima district had over twice as many registered voters.

With the Block Vote system, voters had as many votes as there were seats to be filled within the district, but not all voters made use of all their votes. There was widespread belief that in the 1989 elections voters cast one or two votes for candidates with whom they had family or kinship ties, and then cast subsequent ballots for members of the Muslim Brotherhood, the pre-eminent Islamic political movement. Although the non-party political nature of these elections makes political analysis rather speculative, the University of Jordan has estimated that Muslim Brotherhood candidates won approximately 30 percent of the seats with less than 20 percent of the votes, independent Islamics won 16 percent of the seats, again with far fewer votes, while pro-monarchist candidates won nearly 60 percent of the total vote but only filled 40 percent of seats. These results led King Hussein to believe that the Block Vote gave advantages to Muslim Brotherhood candidates, the most organized and coherent political movement in the embryonic party system, over pro-monarchist independents.

It was for this reason that a new electoral system was introduced by Royal decree for the 1993 general elections; but at the same time Hussein lifted the ban on political parties, and this led to the emergence of a formal Islamic Action Front Party. Believing (probably correctly) that most Jordanian voters felt loyalty to family and kin first and to political ideology second, Hussein decided to maintain the multi-member districts but change the law to one where voters could only choose one candidate in their district. Thus, in a somewhat accidental manner, Jordan adopted the Single Non-Transferable Vote (SNTV), see Single Non-Transferable Vote. In the Jordanian context SNTV is called "one man, one vote", even though this terminology in other countries is primarily used to indicate the fundamental principle of equality between voters rather than a particular electoral system.

In 1993, participation increased slightly from the previous election, but it remained below 50 percent of the voting-age population. The decrease in the number of votes given to individuals forced all electors to consider what was their most important allegiance, political or otherwise. However, what was notable about the Jordanian House of Representatives elected in 1993 was that it contained a much more balanced and representative mix of party representatives and independents than had been previously the case. The Islamic Action Front won 20 percent of the seats with around 17 percent of the votes, Independent pro-monarchists won 60 percent of the seats with 58 percent of the votes, and smaller groupings of independent Islamists, Leftist, Nationalist, and Fateh Movement candidates won a handful of seats with a handful of votes. These results fit in well with the general expectation that SNTV should be much better than the Block Vote in providing a parliament which is relatively proportional to the vote distribution overall - a picture seen in other countries which use or have used SNTV, such as Japan from 1948 to 1995, see Japan - Electoral Reform, and Taiwan.

Nevertheless, the reduction in choice given to voters, combined with the running of a considerable number of Islamic Action Front candidates, led to frustration in a number of quarters over the electoral law changes. During the run-up to the 1997 elections there have been calls to return to the 1989 system of the Block Vote or to adopt a new proportional or mixed electoral system. However, it is likely that Jordan will remain one of only two current examples of an SNTV system (along with Vanuatu) until the end of the century.

Japan: Adapting to a New Electoral System

In 1993 the long-dominant Liberal Democratic Party (LDP) split and lost control of the main chamber of the Japanese Diet in the general election that followed. One of the achievements of the new coalition that formed in its place was reform of the electoral system, which had been widely viewed as a source of corruption and the basis of the LDP’s long-standing dominance.

Under the old electoral system (SNTV), the 511 members of the House of Representatives (the lower house) were elected from 129 districts of between one and six seats each. This system had been in use since 1947 and had produced a distinctive approach to elections among the major parties, particularly the LDP. Under this system any party that hoped to win enough seats to obtain a majority or a significant minority of seats needed to put up multiple candidates in most districts. Thus, in order to maximize their representation, parties needed to find methods of ensuring that each candidate would poll the minimum number of votes required to be elected, rather than having each candidate follow his natural instincts by attempting to maximize his vote. A candidate who received more than his ‘fair share’ of the vote could actually hurt colleagues who received fewer votes: candidate A’s ‘unnecessary votes’ could be enough to prevent candidate B of the same party from gaining a seat.

The LDP dealt with this problem through particularistic policies that targeted selected groups of voters and provided them with ‘pork’ and other benefits. As the first winning party under the SNTV system, the LDP controlled the spoils of office, making it difficult for the various opposition parties to mount an effective challenge. Not surprisingly, this system contributed to corruption. Furthermore, under such a personal and particularistic system, political choice and debate based on substantive policy issues were not given due importance.

By the early 1990s citizens’ anger at the system had produced great pressure for electoral reform. The LDP’s inability to agree on and pass reform legislation contributed to a split in the party that gave power to the opposition (including the LDP defectors) in 1993. The concept of a US-style two-party system and frequent alternation of parties in government had grown in popularity among politicians, scholars and the media, and had come to be seen as a ‘magic bullet’ that would solve the problems of the Japanese political system. As a result, many called for the establishment of a system of single-member districts (SMDs). However, members of the smaller parties in the new government feared that this would crowd them out of the system and thus opposed such a move. The resulting compromise created the two-tier system that is in use today.

The reformed electoral system is a Parallel system consisting of two tiers—List PR and FPTP single-member districts. Each voter casts one vote in each tier. For the first election under this system, in 1996, there were 200 seats in the PR tier divided between 11 regional districts, ranging in size from seven to 33 seats, and 300 SMDs in the second tier. Efforts at rationalization led the Diet to reduce the number of PR seats to 180 prior to the second election in 2000. The 11 PR districts now range in size from six to 29 seats. In a Parallel system, there is no compensatory mechanism that adjusts the overall number of seats won by each party to better reflect the proportion of the vote actually received. The predominance of SMD seats over PR seats thus advantages larger parties that can win SMD seats. The two tiers of the Japanese electoral system are related in another, more unusual, way, however. Japan’s electoral laws allow candidates to mount dual candidacies by standing both on a PR list and for an SMD seat.

While the PR tier is technically closed-list, there is also a provision that allows for some degree of voter influence over the ranking of candidates on the lists. Parties are allowed to present lists that give equal rankings to some or all of those candidates who are nominated both on a party list and for an SMD. After those who win in the SMDs are removed from consideration, the final ranking of the SMD losers on the PR list is determined by how well each polled in comparison to the winner in his or her district.

This provision has a number of benefits for parties. First, it allows them to abdicate the politically challenging job of ranking candidates. Second, it encourages candidates who are ranked equally on the PR lists to campaign more vigorously to win votes in their districts. While parties do make much use of equal ranking, they also retain the option to give some candidates firm rankings. This is also useful, as a higher or ‘safe’ ranking on the PR list can be used as an incentive to convince a candidate to run in a single-member district in which there is little chance of winning.

The first trial of the system came in 1996, and the results were largely seen as disappointing. In the years since the new electoral laws were passed, the LDP had re-established itself in power and the opposition parties had undergone a number of realignments. This instability led to the persistence of previous patterns, an overall win for the LDP, and little movement towards the hoped-for two-party system. The somewhat complicated nature of the system also produced dissatisfaction among the electorate, particularly regarding the phenomenon of losing SMD candidates being ‘resurrected’ in the PR tier. The results were especially counter-intuitive in cases in which the first- and third- (and occasionally fourth-) placed candidates from a single-member district won seats but the second-placed candidate (usually from the most competitive of the opposition parties) failed to win a place. It was also unclear that any significant decline in corruption and money politics had taken place.

By the time of the second election under the new system, in 2000, there had been a reduction in the number of competitive candidates vying for each SMD seat. However, the move towards a two-party system again made only slight progress as the non-communist opposition was still splintered and the centrist Komeito party had switched sides and joined the LDP-led coalition.

The third test of the new system took place in November 2003. In September, the small Liberal Party merged with the dominant opposition Democratic Party (DPJ). The merged party (which retained the DPJ name) gained an impressive 40 seats in an election that featured the use of party manifestos for the first time. The remaining opposition parties of significant size lost all but a few of their seats. On the government side, the LDP and the smaller of its two coalition parties also lost seats, leading to the smaller party being absorbed by the LDP. With most seats concentrated in the hands of the two leading parties, only Komeito remains as a significant small party. The LDP is still in coalition with Komeito, in part because it needs Komeito support in the upper house, but also because support from the well-organized Komeito played a large part in the victories of many of its SMD candidates.

The results of the legislative election of 2003 support the idea that the effects of electoral system reform are not felt immediately and that entrenched habits and processes require time to change. These outcomes also suggest that the mixed-member system may not be likely to produce a complete consolidation into a US-style two-party system, as the existence of the PR tier allows third parties to persist.

Kyrgyzstan: Electoral Manipulation in Central Asia

In contemporary Central Asia, elections are as much political theatre as contests for office. After the break-up of the Soviet Union in late 1991, most of the countries in the region descended into one-man rule or civil war. The semi-competitive elections held in the last months of the Soviet order gave way to elections of acclamation in the first years of independence, with political power becoming increasingly centralized in the hands of the founding presidents of the republics. For a time it appeared that Kyrgyzstan might resist the temptation of authoritarianism; however, by the mid-1990s its president had begun to limit society’s ability to hold the state and its representatives accountable.

The election that brought to power the country’s first president, Askar Akaev, illustrates the role of changing rules in shaping electoral outcomes. In the late Soviet era, parliaments selected the head of state—the chair of the Supreme Soviet—in each republic. In Kyrgyzstan, the election law stipulated that if the Parliament failed to produce a winner after two rounds of voting all the candidates would be disqualified. In October 1990, this quirk in the electoral rules allowed Akaev—a little-regarded Gorbachev loyalist who was opposed to the dominant conservative forces in the Kyrgyz Communist Party—to win the next round of the parliamentary election for head of state of the Kyrgyz Republic. The following year, Kyrgyzstan, like most other Soviet republics, introduced popular direct elections for a newly-designed office of president whose powers supplanted those of the collapsing Communist Party. In October 1991, just weeks before Kyrgyzstan became an independent country, Akaev won the election for the presidency unopposed. He won the two subsequent presidential elections—in December 1995 and October 2000—by wide margins in the first round, although widespread violations were reported during both elections.

The rules governing presidential elections in Kyrgyzstan are a mixture of traditional and unconventional elements. Elections are held every five years and are decided by a two-round majority run-off system: if no candidate receives an absolute majority in the first round, the two candidates with the most votes proceed to a second round, where the candidate with the most votes wins. New elections must be called if less than half the electorate turns out for either the first or the second round. Presidents may serve for no more than two terms, although the Constitutional Court in Kyrgyzstan, unlike its counterpart in the Russian Federation, made an exception for the sitting president by ruling that his first term did not count because it began before the limit of two terms was adopted in the 1993 constitution.

To stand for president, a candidate must be at least 35 and not more than 65 years of age. Candidates must also satisfy several further requirements. First, they must undergo an examination by the Language Commission to ensure that they are fluent in the state language, Kyrgyz. This requirement, introduced to discourage Russians and Russified Kyrgyz from contesting the presidency, was used in the 2000 election to disqualify Akaev’s most prominent challenger, Feliks Kulov. Second, they must pay from their personal funds a deposit equal to 1,000 times the minimum monthly wage—essentially the lifetime income of a poor person. For the deposit to be returned, a candidate must receive 10 per cent of the vote, and proposals now being debated by Parliament would increase that to 15 per cent. A further barrier to entry is the requirement that a candidate receive 50,000 signatures, of which at least 3 per cent must come from each of the country’s eight territories—a provision designed to ensure that a president has adequate support in both the north and the south, whose elites have been at odds in recent years.

The relative stability of the rules governing presidential elections in Kyrgyzstan contrasts with the frequent changes made in the parliamentary electoral system. Perhaps the most dramatic have been to the size and structure of the Parliament. Independent Kyrgyzstan inherited from the Soviet era a unicameral Parliament of 350 deputies who had been elected in February 1990 in single-member districts using a two-round voting system. Following constitutional changes made in 1994 by referendum—the president’s preferred means of enhancing his powers and reducing those of the Parliament—this unicameral assembly was replaced by a bicameral legislature, with 60 members in the Legislative Assembly and 45 in the Assembly of People’s Representatives. In the parliamentary elections of February 1995 and February 2000, the entire Assembly of People’s Representatives and 45 members of the Legislative Assembly were elected in 45 single-member districts using two-round voting. The remaining 15 members of the Legislative Assembly were elected by List PR using closed lists and a single nationwide district with a 5 per cent formal threshold, that is, parties must secure at least 5 per cent of the total vote nationwide to be represented in the Parliament. For the 15 PR seats, each party had the right to put forward a list of 30 persons, and in cases where candidates from the list also stood in single-member districts and won, their names were removed from the party list.

The reduction of the number of deputies from 350 to 105, ostensibly designed as a cost-saving measure, facilitated presidential control of the Parliament by trebling the size of the single-member districts and thus reducing the ability of smaller parties to win seats. The presence of a handful of List PR seats in the new Parliament did little to compensate for the disadvantages that a diminutive Parliament posed for small parties.

Moreover, the post-communist elections have returned parliaments whose composition differed dramatically from that of the rubber-stamp Soviet legislatures. Communist Party control of candidate nomination had worked in such a way as to create bodies in which those who had passed the approval process comprised a broad cross-section of society. In contrast, the post-communist assemblies in Kyrgyzstan were almost exclusively male and had a disproportionate number of executive officials and the newly rich.

Kyrgyzstan has recently changed the rules for parliamentary elections again. Revisions to the constitution adopted by referendum in February 2003 called for the 105-member bicameral assembly to be replaced at the next parliamentary election with a unicameral legislature of 75 members. The new election law of January 2004, which has been much criticized inside and outside Kyrgyzstan, provides that the 75 deputies will be elected in single-member districts using a two-round majority run-off voting system. Further reducing the size of the Assembly and abandoning the party list seats is likely to reduce the representation of minorities yet again, increase the executive branch’s influence over the legislature and emasculate an already weak party system. It may also strengthen the political salience of the regions by giving the central party leaders less influence over the selection of candidates.

Because the smaller number of seats in recent parliaments produced larger electoral districts, it has been easier for ethnic Kyrgyz to win seats than for members of ethnic minorities. Where the ethnic Kyrgyz majority is now over-represented in the Parliament, the substantial Uzbek, Russian and German minorities are all significantly under-represented. In particular, the Uzbeks hold a share of the seats which is less than half of their share of the population.

In recent years, the political opposition in Kyrgyzstan has found it increasingly difficult to contest presidential and parliamentary elections. The deference of the judiciary, the Electoral Commission and the Language Commission to presidential authority has led to the selective prosecution and disqualification of electoral candidates. Moreover, presidential influence on the media has prevented the opposition from waging effective campaigns. In the 2000 presidential election, for example, President Akaev received almost ten hours of coverage on the national television channel, KTR, while his principal opponent received less than five minutes. One of the few sources of independent reporting on electoral campaigns, the foreign press, is threatened with legal sanctions if it criticizes establishment candidates. Voting irregularities are also widespread. The conduct of elections as well as the changing electoral rules has impeded the development of political competition in Kyrgyzstan.

For most of the first decade of independence, elections to representative assemblies below the national level were held in single-member districts using a two-round voting system. Since 1999, however, regional and local assembly elections have been conducted in multi-member districts using SNTV. Although the governors of the country’s seven regions are still appointed by the president, the chief executives of cities, districts and villages are now selected by the members of the local assemblies. The sole exception to this pattern is the capital, Bishkek, where the mayor is directly elected.

As in Georgia and Ukraine, the manipulation of electoral rules and the conduct of elections ultimately delegitimized the elections themselves, which contributed to the March 24, 2005 revolution in Kyrgyzstan that overthrew the Akaev presidency and placed the newly elected parliament and the entire system of electoral rules under review.

Sri Lanka: Changes to Accommodate Diversity

Like other former British colonies, Sri Lanka inherited a Westminster model of parliamentary government, with universal suffrage established in 1931 and full general elections in 1947, but over time found that First Past the Post (FPTP) elections were incapable of representing minority interests, see First Past the Post (FPTP). In 1978, the decision was taken to transform Sri Lankan government from a parliamentary system into a French-style executive presidency, and a Select Committee was appointed to consider the necessary wide-ranging constitutional changes.

Sri Lanka is a nation with a long history of bitter ethnic conflict between the majority Sinhalese and minority Tamil communities. It was for this reason that the constitutional drafters were very conscious of the need to ensure that the new office of executive president would be filled by a national figure representative of all groups in society, and capable of encouraging consensual politics between those groups. The new president would have to represent all groups in Sri Lankan society and be seen as a figure capable of moderating between opposing interests. These objectives focussed attention on the method of election to the new office, and particularly on the means by which ethnic minorities could be included in the selection process.

While the method of election for such a figure would be crucial to the fate of the office, and would require at least a majority of voters supporting the successful candidate, only once in 50 years had any political party secured a majority of the vote at a national election; indeed, most governments had been elected with considerably less than that. The party system in Sri Lanka was fragmented between two dominant Sinhalese parties and a number of small minority parties.

Because much of the 1978 constitution had its philosophic origins in the French Fifth Republic model of a strong executive presidency combined with an elected legislature, the initial plans provided for a Two-Round System (TRS) of presidential elections, see Two-Round System. However, the extra cost and security issues associated with holding two separate elections within a two-week period was seen as being a major defect, particularly since Sri Lanka was in the midst of a violent civil war at the time.

These considerations prompted an innovative solution to the problem: to combine the initial and run-off rounds of voting into one election via the expression of preferences. Under this system, which continues to be used, if no candidate has a majority of first preferences, all candidates other than the two leaders are eliminated, and the second preferences distributed to one or the other of the top two to ensure a majority winner; voters can number up to three preferences, which will then be distributed to one or the other of the top two candidates in the event of no candidate having an absolute majority. The system thus achieves in one election what a Two-Round System achieves in two, see Alternative Vote.

In addition to ensuring that the president would be elected, whether outright or via preferences, by an absolute majority of all voters, the system has the additional feature of encouraging candidates to look beyond their own party or ethnic group for second-preference support from other groups. Sri Lanka has now conducted three national presidential elections under the supplementary vote system, in 1982, 1988, and 1994. Contrary to expectations, at each of these elections the winning candidate has achieved an absolute majority in the first round, and thus no preferences have been counted. The possibility that preferences may one day decide the result, however, does appear to have influenced the campaign strategies of Sri Lankan parties, and there is considerably more attention paid to minority groups in election campaigning for presidential elections than was formerly the case.

Lesotho: Africa’s First MMP Electoral System

The result of the ordinary National Assembly elections in Lesotho in May 1998 was very clear. The governing Lesotho Congress for Democracy (LCD) won an overwhelming electoral victory, taking 79 of the 80 seats in the National Assembly.

The only problem was that LCD had only received the support of slightly over 60 per cent of the electorate. The result in terms of number of seats won was yet another example of how the FPTP electoral system can lead to remarkable discrepancies between the share of the vote and the share of the seats won by political parties. A discrepancy of this kind should not come as a surprise–it had happened before–but it was followed by the losing parties, and especially the main opposition party, the Basotho National Party (BNP), crying ‘Foul’. This was also nothing new, but it was a sad surprise that the accusations about the overall correctness of the 1998 election results (which were never seriously challenged), some time after they were published, incensed the public to such a degree that they started rioting in the streets of the capital, Maseru, setting fire to and demolishing public as well as private buildings.

The government called on the Southern African Development Community (SADC) to intervene, and the SADC did so, relying mainly on the armed forces of South Africa. When order was restored, an agreement was reached on 2 October 1998 (later guaranteed by the SADC) which called for the establishment of an Interim Political Authority (IPA) on which the 12 parties which had put up candidates in the elections were given two seats each, no matter how small their electoral support. The IPA’s brief was to develop a new electoral system and suggest other political and administrative measures to strengthen Lesotho’s peaceful democratic development. However, all recommendations would be submitted to the government, which would then take them to Parliament to be enacted in the ordinary way.

The ideas behind the establishment of the IPA were clearly inspired by the institutions of the negotiation process in South Africa during the early part of the transition process, but it was not taken into account that the two processes were so different that the institutional solutions also had to be different. The subsequent political process in Lesotho was not an easy one, and it can be no surprise that the opposition’s overwhelming majority on the IPA—by 22 to 2 against the government of the day—was not conducive to a constructive climate of negotiation.

The IPA representatives, none of whom had been able to ensure a seat for themselves in the National Assembly, were eager to suggest an electoral system which would keep the single-member constituencies and at the same time provide for a much more proportional outcome at the next elections than had been the case in 1998. The obvious solution was either an MMP or a Parallel System. A German expert on electoral systems was invited to give a presentation, after which the IPA majority opted for the MMP solution, with some seats allocated in single-member districts and others allocated from party lists on a compensatory basis. The LCD—in complete control of the legislature—decided to opt for the alternative, the Parallel System, which would give it, on top of its expected massive share of the single-member district seats, an additional number reflecting its share of the votes cast for the seats not allocated in the single-member districts.

It soon transpired that the IPA was not aware of all the practical details that should be taken into consideration when deciding to go for MMP, such as the seat allocation formula, the issue of a formal electoral threshold, overhang mandates, one or two ballots, and so on. The number of seats in the two categories was also an issue, even though most IPA members seemed to agree that keeping the 80 single-member districts was a good idea and that it was only natural to have 50 compensatory seats. The basis for the latter suggestion was a little awkward: previously Lesotho had had 65 single-member districts. If it returned to that number and then added an identical number of compensatory seats (as in Germany), the National Assembly would have 130 seats altogether. However, if the size of the Assembly was to be 130, as the number of single-member districts for the time being was 80, and as it was difficult to imagine this number being changed in the immediate future, the number of compensatory seats had to be 50. The government challenged this number, among other reasons because Lesotho is a small and poor country which should only have a reasonable number of parliamentary seats.

The political conflict was easy to understand. The IPA, which was tasked with suggesting institutional solutions to the political impasse, was strongly in favour of MMP with 80 single-member districts and 50 compensatory seats, while the government—in complete control of the legislature, which had to pass all the IPA’s suggestions—argued that the best solution was a Parallel system with the same 80 single-member districts and probably 40 seats to be allocated separately on the basis of (preferably) the same ballot as was used in the single-member districts, although a second ballot was also an option.

The political compromise over the electoral system took some time to reach, mainly because of the level of distrust between the two sides and some hesitation about the very idea of compromise. It was eventually agreed that the electoral system should be MMP (which was the main opposition objective), while the number of seats should be 120 (80 + 40), which was very important for the government side. While the government held all the cards through its huge parliamentary majority, it was clear that some concession had to be made in order to ensure wider acceptance and thus legitimacy of the revisions. The consequent constitutional amendment required strong support not only in the National Assembly but also in the Senate (made up mainly of chiefs), which was another reason why compromise was necessary. The reason for this was that, if the two houses of Parliament could not agree on the constitutional amendment, it had to be put to a popular referendum, and this was not really possible because of disagreement over the electoral register. Eventually, the constitutional amendment was formally adopted in May 2001. Only then could the necessary changes to the electoral law be considered.

The 1998 internationally guaranteed agreement had provided for early elections, to take place in May 2000. This was completely unrealistic, not least because the government and the opposition (the IPA) were not really on speaking terms, and a new Independent Electoral Commission had only been appointed in April. Agreement was then reached on delaying the election by a year, but further delays in reaching agreement about the electoral system, concerns about an adequate voter registration system and so on meant that a new general election was only possible in May 2002.

The election went reasonably well. The LCD, not entirely unexpectedly, won 55 per cent of the party (PR) votes but 65 per cent of all the seats. The reason for this was that the party won 77 of the 78 single-member districts contested on election day (elections in the remaining two were postponed because of the death of candidates, but eventually the LCD also won them). The system does not have overhang mandates, so the opposition got all 40 compensatory seats.

Seven of the eight opposition parties which won seats ended up being under-represented in terms of share of votes compared to share of seats. This under-representation was, however, very much smaller than it had been in 1998, and the National Assembly of Lesotho is now a fairly representative body in terms of political representativeness. Thus the main objectives of the efforts after the 1998 troubles have certainly been achieved.

It is clear that the combination of (a) one party taking almost all the single-member districts, (b) only 33 per cent of the seats being compensatory seats, and (c) the absence of surplus seats may continue to cause some degree of disproportionality in future elections as well. However, this seems a small price to pay for the various improvements in the system achieved during the protracted political compromise-seeking process of 1999–2001, when it was also a concern not to have too many members of Parliament in a small and poor country.

Mali: A Two-Round System in Africa

The former French colony of Mali in West Africa made a successful transition to multi-party politics in 1991, after three decades of authoritarian rule. Principal among the new democratic institutions established at the time was a 129-seat National Assembly, with 116 seats elected by the domestic electorate and 13 by Malians residing overseas. The 116 domestic seats are allocated on the basis of population (one seat per 60,000 people) among 55 constituencies (circonscriptions) corresponding to the country's 49 administrative divisions (cercles) and the six communes in Bamako, the capital. Because of population disparities, the district magnitudes range from one to six seats per constituency.

While independent candidates are permitted, political parties are required to submit closed-party lists with the same number of candidates as available seats, see List PR. Voters exercise their choice through categorical ballots, so they can vote for only one independent or party list of candidates. A Two-Round majority-runoff system is used whereby, in the absence of an independent candidate or party list winning an absolute majority in the first round, only the top two finishers in the first round compete in the second round, with the winner decided by an absolute majority, see The Two-Round System. In the case of the multi-member districts, the two highest party lists from the first round compete in the second, with the winning list gaining every seat in the district. A similar Two-Round majority formula is used in the presidential election. A proportional representation formula based on the largest remainder - the Hare quota - is used in municipal elections.

As in most of francophone Africa, the new democratic institutions in Mali were debated and selected in a broadly-based National Conference, which included three representatives of each of the officially-registered political parties. The electoral system that emerged out of this process was a compromise aimed both at preserving the political power of the five major parties while creating electoral opportunities for numerous smaller ones, and at balancing the contradictory imperatives of securing broad political representation and producing stable governing majorities. Thus the initial proposal for using the Two-Round System (TRS) in single-member districts was rejected, in order to diminish the influence of local notables and strengthen party control over candidates. Also rejected was a proposal from smaller parties for a PR system, because of its anticipated potential for political instability. However, the adoption of the PR formula for municipal elections accommodated the smaller parties, most of which lacked national support and was regionally or locally based. Conversely, it was thought that a Two-Round majority-runoff system for legislative elections would encourage coalitions in the second round between smaller and larger parties. The adoption of the Two-Round majority-runoff formula for presidential elections reflects the consensus in most African countries that the head of the state must be supported by a majority of the electorate.

The new Malian electoral system produced a relatively fair and competitive electoral process in 1992. The first round was contested by 23 officially registered parties, including three with national political bases, Alliance pour la democratie au Mali (ADEMA), Congres national d'Initiative démocratique (CNID), and Union soudanaise-Rassemblement démocratique africaine (USRDA), and two with a limited national base but with the potential of becoming national parties, Rassemblement pour la democratie et le progres (RDP) and Parti progressiste soudanaise (PSP). The rest had regional and local bases and no prospect of entering the national government without forming a coalition with the other five. The competitiveness of the system was illustrated by the fact that only 11 out of 44 constituencies were decided in the first round, with 15 seats won by the five parties. Of the ten contending parties in the second round, six had led in at least one constituency, but the leading party list was defeated in seven of the 44 constituencies. Indeed, each of the five major parties lost second-round district elections after leading in the first round.

Combined with the entry of large numbers of small parties with limited electoral support, a phenomenon that is typical of new democracies established after an extended period of authoritarian rule, Mali's new system produced the expected political impact on vote-seat disproportionality and multi-partism. Thus, the Two-Round majority formula produced a high level of disproportionality (between seats and votes), a moderate degree of electoral multi-partism (3.3 effective electoral parties), and a moderately low legislative multi-partism (2.2 effective legislative parties).

The Malian electoral system has effectively balanced representation and governance, but at the same time fostered a viable parliamentary opposition. The use of closed party lists in multi-member districts, moreover, has encouraged ethnic and regional alliances among otherwise socially fragmented and politically weak groups. However, several problems remain. First, the very use of party lists weakens the constituency linkages of elected representatives. Confronted with strong pressures for constituency work, many MPs have informally divided up their constituencies into individual bailiwicks for that purpose. Second, the National Assembly possesses only a limited capacity to check executive authority, since institutionally its powers remain weak relative to the strong executive presidency. And finally, this problem is compounded by the disproportionately large percentage (66 percent) of seats won by ADEMA, the incumbent ruling party, due in part to the electoral formula and in part to population disparities among the constituencies, and especially in rural constituencies.

These problems prompted opposition demands for electoral reform. This led to political negotiations between the opposition parties and ADEMA which produced agreements on three issues prior to the April 1997 legislative elections: the use of PR formulae for allocating some National Assembly seats, which was subsequently declared unconstitutional by the judiciary, a 27 percent increase in the size of the National Assembly from 116 to 147, with a reduction in single-member and a corresponding increase in multi-member constituencies, potentially giving the opposition parties a degree of electoral advantage, and the creation of a broadly representative Electoral Commission. However, the quickly created Commission was unprepared to take on the complex task of election management. The ensuing logistical and administrative problems provoked opposition demands for the annulment of the 1997 legislative elections, to which ADEMA agreed, even though early returns confirmed predictions about its victory.

The fact that these changes in the electoral system were negotiated attests to the success of Mali's new democracy in managing political conflicts peacefully. It also indicates that the choice and reform of new democratic institutions are not pre-determined, but are negotiated outcomes of which future political consequences are often obscure. To what extent the recent reforms of Mali's electoral system will have the desired effect when they are implemented remains to be seen.

Malta: STV With Some Twists

Single-Transferable Vote (STV) has been in use in Maltese elections since 1921, long before this small Mediterranean Island nation achieved independence from Britain in 1964. Although Malta subsequently became a republic and replaced the office of Governor-General (representative of the Queen) with a President, it retained the Westminster model of parliamentary democracy. The constitution mandates election of the members of the House of Representatives, Malta's unicameral parliament, "upon the principle of proportional representation by means of the single transferable vote." The maximum length of a parliamentary term is five years, but the legislature may be dissolved earlier. There are no other elective offices except for local councils introduced by the 1993 Local Councils Act, whose members are elected by STV as well.

For purposes of parliamentary election the country is currently divided into 13 divisions, all of which is of roughly the same population size. In contrast to Ireland, each district elects the same number of MPs (five), for a total size of parliament of 65 (ignoring any bonus seats). Each seat corresponds to approximately 4,200 registered voters. In 1996 the quota needed to secure a seat ranged from 3,245 to 3,519 votes. Candidates may simultaneously stand in two divisions. If they win seats in both, they must resign one, which is then filled through a so-called "casual election". This is not a by-election in the conventional sense; instead the winner is determined by applying STV procedures to the ballot papers credited to the vacating candidate in the general election.

On the ballot, candidates are listed alphabetically within party blocs. Voters express their preferences by placing sequential numbers next to the candidates' names. There is no obligation to rank-order all of them or to stick to candidates of a single party. Indeed only a single preference (indicated by the number 1) is required for a vote to be valid. Unlike their Australian counterparts, Maltese voters do not have a whole-ticket option. Nor do the parties prepare a recommended rank-order of candidates.

There are three noteworthy characteristics of Malta's experience with STV:

The first is that although STV can function as a nonpartisan election method, partisanship is a prominent feature of electoral contests in Malta. Voters enjoy but make little use of the opportunity to cross party lines when ranking candidates on their ballots. For this reason a minute percentage of votes (one percent) transfers to candidates of other parties.
A second particularity is the practice of the two major political parties to nominate many more candidates than could possibly win in a district. This may at least in part be explained by the loyalty pattern in preference voting. The parties apparently do not fear a loss of votes due to over-nomination because preference votes given to their less popular candidates will ultimately transfer to other candidates of their party. At the same time, a larger and more variegated roster of candidates may help them attract more votes. For the candidates, of course, this means that they face very intense competition from within the ranks of their own party and must go to great length to earn and retain voter support. To win a seat and to keep it, a politician has to build and maintain a personal support base, but since the vote is secret and the supporters in the constituency are not identified, he or she is well-advised to appeal to and serve a much larger group. This produces very close relationships between representatives and their constituents. The voters have the benefit of being able to call on several MPs representing their district. Due to the fact that at least one member of each major party is elected from each district, they even have a choice by party.
The third important particularity is that modern Malta has a virtually pure two-party system. Indeed this is unusual for PR systems, which reduce barriers for small political parties. The reasons why third parties, which do exist in Malta, have failed to thrive electorally in recent decades are not entirely clear. But the implications of this situation are important: If MPs of only two parties are elected to an odd-number-sized parliament, then one of them will necessarily command a majority and form the government. Moreover, the two major parties, the Maltese Labour Party (MLP) and Nationalist Party (PN), enjoy nearly equal support in the electorate and are thus very competitive. This means that even small distortions in the vote-seat ratios can drastically affect the outcome of an election and thus the control of government. This has in fact been one of the most severe problems with STV in Malta.

In 1981 the MLP won a majority of seats in parliament even though Nationalist Party candidates had received a majority of first-preference votes nation-wide. Allegedly this had occurred as a result of deliberate gerrymandering by the MLP government, although such charges are hard to prove. More importantly, however, this seemingly "perverse" result led to a major constitutional crisis when the Nationalist refused to accept the outcome of the election and walked out of Parliament, thus putting the legitimacy of the entire system in doubt.

The Nationalist boycott ended when the MLP agreed to discuss constitutional reforms to prevent a recurrence of the scenario of victors being turned into losers. In 1987 the constitution was changed accordingly. Article 52, as amended, assures that the party with a majority of first-preference votes will receive as many additional seats as necessary to give it a majority in parliament, thus allowing it to form the next government. A second amendment, adopted more recently, provides for a similar adjustment for the party with the most votes (but not a majority) where more than two parties compete for votes but only two parties win seats in Parliament. In 1987 and 1996 additional seats were thus allocated to the Nationalist Party and the MLP, respectively.

Because of these constitutional amendments the voters' first preference on the ballot is now not only an ordinal vote for the most-favored candidate, but also a categorical vote for a party. In addition, at least as long as the two-party system perseveres, it is an expression of a preference on which party shall form the government. A general election can thus be said to provide a clear judgment on the record of the incumbent government and a clear mandate for the victorious party.

What lessons might be derived from Malta's experience with STV?

Malta can be said to provide a cautionary tale. While useful generalizations can be made about the effects of electoral systems, there are sometimes unique circumstances that lead to unexpected results. As seen here, a highly proportional electoral system is also subject to failure under certain conditions. At the same time, however, Malta's handling of the ensuing crisis is cause for optimism, for it provides an apt illustration of how constitutional engineering solutions can be found to redress institutional failures when they occur, and how they can be implemented through bargaining and compromise.

Leaving aside the disproportionality issue that came to a head in 1981, we must note that Malta has had a series of single-party governments and a fair amount of alternation in partisan control. The intense intra-party competition engendered by STV combined with over-nomination has not had the effect of rendering the parties ineffective as political organizations either in government or in opposition.

Malta's experience in administering the allegedly complex single transferable vote system is also encouraging. Although the determination of winners is more cumbersome and time-consuming than is the case with other systems, the process is manageable. The number of counts necessary to fill all seats in a district is not a function of the number of voters/ballots, but of the number of candidates in that district, although a larger electorate (and thus larger number of ballots) will of course increase the workload. Nor do Maltese voters appear to be overly perplexed by their system. Voter participation is almost universal (more than 95 percent in recent elections) and the percentage of invalid ballots is low (rarely more than one percent).

Like the Republic of Ireland, Malta is not an ethnically or religiously diverse country, and thus provides no opportunity to assess the performance of STV in terms of minority representation. It is clear, however, that under Malta's version of STV, minorities would be assured of their ability to elect candidates of their choice irrespective of the preferences of the majority as long as their members make up 17 percent of the voters in any district [Quota = votes/(5+1) +1]. Increasing the number of seats per district could lower this threshold further, although there are obvious practical limitations in terms of ballot length and complexity. It is also clear that women are in a position to elect at least two MPs in each district (or 40 percent of the seats) regardless of the voting preferences of men, although this potential voting power does not currently translate into the election of large numbers of female candidates in Malta.

STV has many favorable characteristics in theory and has worked well in Ireland, Malta, and Australia in practice. What is lacking is a broader experiential base to learn from since the actual use of the system is limited to the Anglo-American world (with very few exceptions). We do not know for sure how it would perform in a variety of other settings. One thing is hardly in doubt, however: STV hands voters the most sophisticated instrument to express their preferences; it meticulously aggregates these diverse preferences and translates them into parliamentary representation. Even where parties are as strong and predominant in politics as in Malta, STV still assures that the voting public will determine the identity of all the politicians that take up seats in parliament to collectively represent the will of the people. Where such grass-roots democratic control is deemed desirable, STV would seem to be the system of choice. The flip side is that the ability of the party leadership to determine the composition of its parliamentary group is limited correspondingly.

Mexico: Democratization Through Electoral Reform

Mexico has a presidential system with strong and independent legislative, executive and judicial branches. The doctrine of the separation of powers, which did not function in practice between 1929 and 1997, when the single official party, the Institutional Revolutionary Party (PRI), controlled both the executive and Congress, has been resurrected and is now the dominant feature of politics at the federal level.

The president is elected by plurality vote. In the 1988 and 1994 elections, the winner won about half of the votes cast, but in the 2000 election the winner, Vicente Fox, won only 42.5 per cent of the votes. Proposals exist to amend the constitution to introduce a run-off election between the two front-runners if no candidate wins an absolute majority in the first round. Their success will depend primarily on the electoral prospects of the major parties, as well as considerations of the cost of a second round.

The president is elected for a six-year term and can never be re-elected or reappointed. This prevents presidents from becoming entrenched in power, but it also diminishes their accountability because they never have to face the electorate again. Considering the ideological and symbolic roots behind the prohibition on presidential re-election (it was a focal point in the Mexican Revolution), it is unlikely that this clause will be repealed soon.

The Mexican Congress is bicameral, the Chamber of Deputies elected for three-year terms and the Senate elected for six years (synchronized with the presidential term). Both chambers are elected through mixed systems, using FPTP and List PR.

The Chamber of Deputies has 500 members, 300 elected by FPTP in single-member districts (SMDs) and 200 elected by List PR in five 40-member regional districts. The 300 FPTP seats are apportioned to the states in proportion to population, with the restriction that no state can have fewer than two seats. The Federal Electoral Institute (IFE), the independent electoral authority, uses the pure Sainte-Laguë Method to allocate seats among the states. The IFE then creates SMDs of roughly equal population within each of the states, generally favouring following municipal boundaries over achieving electoral districts with equal populations, and also divides the country into the five 40-member districts for the purpose of elections to the List PR seats. Each party nominates a candidate for each SMD and presents a rank-ordered list of 40 candidates for each of the five regional districts.

Parties may form total or partial coalitions for electoral purposes, running the same candidate in some districts or sharing PR lists. If they do they must submit agreements to the IFE specifying how the votes in the coalition are to be allocated. If parties form a coalition to elect the president, then they must form a coalition for all the Chamber of Deputies and Senate contests as well. In the 2000 election, two of the three presidential candidates were backed by coalitions. In the 2003 legislative elections, there was a partial coalition between the PRI and the Greens, which ran together in 97 single-member districts and separately in 203, and had separate PR lists (the parties had agreed on how to divide the votes from the 97 districts for the purposes of assigning seats to the List PR candidates).

Voters cast a single ballot for deputies. The sum of all of the votes from the district FPTP contests is then used to calculate the number of PR seats to be allocated to each party, using the Largest Remainder Method and the Hare Quota, and there is a 2 per cent threshold based on the total national vote included in the law. The number of PR seats assigned to a party is independent of the number of FPTP districts won, with two important exceptions: no party can ever win more than 300 seats, and no party’s share of the 500 seats can be more than 8 percentage points higher than its share of the valid vote. A party must therefore win at least 42.2 per cent of the valid vote plus at least 167 districts to win 251 seats in the lower chamber. In 1997 and 2003, the PRI’s share of the seats was limited by the 8 per cent rule. In 2000, the 8 per cent rule did not affect either the PRI or the National Action Party (PAN).

Seats are assigned to party list deputies in the five 40-member regional districts, also using the Hare Quota with largest remainders. The lists are rank-ordered and closed, so that the deputies higher on the list are elected first, and voters cannot modify the order of the list.

The move towards pluralism and multiparty politics in Mexico has been a slow process of evolution. Since 1979 there have been extensive reforms to the electoral formulas used to elect the Chamber of Deputies. The formula used in the 1979, 1982 and 1985 elections had 300 SMDs and 100 party list seats, which were restricted to parties that did not win more than 60 districts. The formula used in 1988 increased the number of party list seats to 200, but guaranteed that the party that won a plurality of districts would win a majority of seats, regardless of its share of the vote. A ceiling was established to the number of seats a single party could win, at 350 seats. The 1991 reforms maintained the ceiling and the majority-assuring clause, but required that the winning party win at least 30 per cent of the vote. It also created bonus seats for the winning party, so that it would not have to function with only a narrow majority in the Chamber. In return, the government ceded some control over the electoral process to a partially autonomous electoral management body (the IFE) and to a federal electoral court. The 1994 reforms eliminated the majority-assuring clause and created a Parallel system, in which the elections to the List PR seats were completely decoupled from the elections to the plurality seats. No party could win more than 60 per cent of the seats (300 of 500) in most circumstances. However, this led to the most disproportional result that Mexico has experienced under mixed systems, with the PRI winning 60 per cent of the seats with about 50 per cent of the vote. So in 1996 the electoral law was adjusted again to set the limit to the number of seats a party could win at 300 and the maximum level of over-representation, as described above, at 8 percentage points. This electoral rule has been as stable as any since multiparty representation was established in 1964, having been used in the 1997, 2000, and 2003 elections. No party has won an absolute majority of seats under this rule. The 1996 reform also made the IFE fully autonomous and enhanced the powers of the federal electoral court. Currently there are proposals to make the Chamber of Deputies either more or less proportional, decreasing or increasing the proportion of list deputies, and decreasing or eliminating the margin of over-representation. However, since no two parties have similar goals, reforms are unlikely to come about. The Senate before 1994 had 64 members, two for each of the 31 states plus the Federal District. The senators were elected under various plurality rules. The result was that until 1988 all senators were members of the PRI. The PRI monopoly in the Senate allowed the government to make concessions to opposition, making the Chamber of Deputies more proportional.

By 1994, there were calls for the Senate to be made more widely representative as well. It was expanded to 128 members, with at least a quarter of the seats guaranteed to the opposition. For the 1997 election, a mixed system was established. Each state elects three senators, and in addition 32 are elected by PR on a single national list. In each state, a party nominates a ranked slate of two Senate candidates. Both candidates of the party that wins the most votes are elected as senators, and the first listed candidate of the party that is placed second wins the third Senate seat. Voters cannot adjust the order of the candidates. Each party also nominates a closed, ranked list of 32 candidates for the national PR list. All the votes for the Senate in each state are totalled at the national level. The formula used is a Largest Remainder Method using the Hare Quota and a 2 per cent threshold. Unlike the Chamber of Deputies, there is no linkage between the plurality and the PR seats; instead, the two systems run in parallel and the PR seats do not compensate for any disproportionality. This electoral formula would create a majority for the largest party if it wins around 40 per cent of the national vote, favourably distributed, and has a margin of three or four points over its nearest rival. Winning two-thirds of the seats in the Senate (important for constitutional reforms, electing Supreme Court justices, and internal procedural matters) requires two-thirds of the national vote. No party won an absolute majority of Senate seats in the 2000 election.

Several proposals have been submitted in Congress to eliminate the party list senators, with arguments that a national list is not appropriate for a chamber that represents the states. However, simply eliminating the PR list would benefit the PRI, which is placed either first or second in all but one of the states, and is thus likely to be opposed by other parties. Alternatives would have three or four senators per state, all elected by PR, most likely using the D’Hondt Formula.

Re-election for consecutive terms is prohibited for all federal deputies and senators (and also for governors, state legislators, mayors, and municipal councillors). Legislators can be elected to the other chamber when their term expires, and they can be re-elected to the same chamber after sitting out a term. The ‘no re-election’ reforms were implemented in 1932 to resolve problems in the PRI by increasing loyalty to the central committee and reducing the power of local party bosses. At the time, the reform was sold as the natural conclusion of the ideology of no re-election from the Mexican Revolution. However, it has served to reduce the autonomy of legislators, because their career prospects after their term of office depended on the party machinery, and for many years increased the power of the president because of his control over his party’s machinery. Party discipline has thus been traditionally very high, approaching 100 per cent for the federal legislators of the PRI up to 2000. This has had profound effects on accountability and representation. Voters can neither reward good performance nor punish poor representation.

All the parties use relatively closed procedures to select candidates—elite designation, closed conventions, or closed or highly controlled primaries. In general, nominating procedures have been opening up in recent years, but candidates are still highly dependent on parties. Additionally, parties control most campaign expenditures, even in district and state contests, and closed lists reduce the incentive for candidates to campaign.

Mexico’s slow democratization has seen frequent electoral system change as a series of concessions by the dominant party to defuse dissent, which has resulted finally in a multiparty presidential system with very strong parties. Further change may now be less likely, as different parties have different interests and any change is seen as a zero-sum game.

Namibia - National List PR in Southern Africa

For much of the twentieth century Namibia was the "forgotten colony" of southern Africa, first occupied by the Germans in 1884, then liberated by South African and allied forces in 1915, only to be effectively recolonized by South Africa who were given trusteeship of the region by the League of Nations in 1920. Apartheid South Africa effectively superimposed her ethnically divisive and exclusionary legal structures upon Namibia (or South-West Africa as it was then known) in the post-war period right up until full independence in 1990. 1989 proved to be a year of rapid change after seventy years of internal struggle and international ambivalence, and years of foot dragging by the South African government who, ten years before in 1978 had agreed to a United Nations (UN) plan for a military withdrawal and transition to independence.

For her "liberation" election of November 1989 and her second parliamentary general election held in December 1994 Namibia used the most basic form of national list Proportional Representation (PR) with the whole country constituting a single district returning 72 members of parliament, see List PR. The allotment of seats was done by the Hare method which, along with the Droop quota, is a largest remainder method but uses a slightly different quota which on average gives more proportional results. No threshold for representation was imposed so the quota became 1.39 percent of the national valid vote but with the largest remainder method the Namibia National Front (NNF) managed to win a single seat with 0.8 percent of the vote in 1989, and the Democratic Coalition of Namibia (DCN) and Monitor Action Group (MAG) won seats with 0.82 percent and 0.83 percent respectively in 1994.

The adoption of list PR came primarily at the instigation of the United Nations who urged as early as 1982 that any future non-racial electoral system ensure that political parties managing to gain substantial support in the election be rewarded with "fair representation." Indeed the Namibians had their new constitution largely imposed upon them. The option of discarding the first past the post electoral system (the whites-only system operating in what was the colony of South-West Africa) and moving to a rigid list PR system was originally proposed by Pik Botha, the then South African Foreign Minister. Although the South Africans had previously, but unsuccessfully, pressed for separate voter rolls (a la Zimbabwe 1980-1985) which would have ensured that whites gain seats in the new Assembly. Botha's subsequent PR proposal was accepted in principle by the UN Secretary-General who then handed over the specific details to the South African Administrator-General and UN Special Representative. The PR system dovetailed nicely with the UN's earlier pronouncements, which stressed the need for as wide as possible representation in the forthcoming Namibian Constituent Assembly.

For the first elections in 1989 the South West African People's Organization (SWAPO) had expressed a preference for keeping the single member district system, no doubt reasonably expecting (as the dominant party) to be advantaged by such winner-take-all constituencies. However, when the Constituent Assembly met for the first time in November 1989, and each parliamentary party presented their draft constitution, SWAPO readily gave in on the issue of PR apparently as a concession to the minority parties for which they hoped to gain reciprocal concessions on matters of more importance.

1989 Election Results

Precursing the subsequent South African general election of 1994 the first multiparty Namibian election of 1989 produced a what many international observers felt to be a "dream result" with the liberation movement (SWAPO) winning handsomely, with 57 percent of the national vote, but not winning enough seats (48) to write the new constitution alone. The opposition, led by the DTA, were in the eyes of many, suitably rejected for their tainted pasts and explicitly ethnic appeals, but they still commanded enough votes and seats to mount a serious opposition within parliament and balance out the possible excesses of SWAPO majority rule. This "positive power configuration in the Constituent Assembly conducive to real compromise" led to a new constitution, adopted in March 1990 which was widely acclaimed as one if the most democratic and enlightened constitutions to be found anywhere in the world.

While SWAPO proved to be the only party capable of winning an absolute majority in the 1989 Namibian elections the wildly inconsistent distribution of their vote led some African political scientists to argue that they had dramatically failed to live up to their promise of being the "sole and authentic representative of the people of Namibia." SWAPO gained over 90 percent of the votes in the large northern electoral district (ED) of Ovamboland, which contributed nearly 60 percent of their national total. Furthermore, while country wide they did manage to out poll all other parties combined SWAPO actually won a majority in only seven, of the 23, electoral districts, with the DTA winning majorities in 15 districts. Similar regional concentrations of party support existed for the minority parties. Besides the DTA, which gleaned most of their votes from the south of the country and other farming areas in Koakoland and Hereroland, the UDF and ACN (the only other parties to win more than one parliamentary seat) drew their support from clearly defined geographical areas and ethnic groups. The UDF was strongly identified with the Damara ethnic group and as expected polled an absolute majority of the votes in Damaraland. While the ACN, a predominantly white grouping, polled nearly half its entire national vote in the EDs of Karasburg (the southern border area where many South African whites voted), and Windhoek (the capital which again played host to a disproportionate number of white voters).

1994 Election Results

The 1994 elections illustrated a consolidation of SWAPO support juxtaposed against severe erosion of the opposition vote. SWAPO actually polled 23,000 votes less in 1994 than they had gained in the first parliamentary elections of 1989 but the much lower turnout (down from 682,000 to 497,000) meant that their share of the vote was pushed up nearly 20 percent, giving them 12 extra seats and more than the two-thirds parliamentary majority needed to change the constitution unilaterally. The stayaway of voters who had cast ballots in 1989 came primarily at the expense of the DTA. Their 101,000 national votes were only just over half the 191,000 they had received in 1989 and translated into a loss of six parliamentary seats. Similarly the UDF lost two of its four seats (retaining only 35 percent of its 1989 vote) while Moses Katjiongua's NPF, now renamed the Democratic Coalition of Namibia (DCN), could do no more than retain its single seat.

SWAPO's overwhelming victory prompted the DTA leader and presidential candidate, Mushake Muyongo, to claim that the election had shown the country to be divided along ethnic lines, and that it had become an ethnic democracy. For the 1994 elections, the Namibian boundary commission discarded the old, uneven SouthWest African electoral districts in favor of 13 new regional districts broken down into 95 smaller counting areas. SWAPO managed to make more inroads into the votes of non-Ovambo communities in 1994 than they had done in the first multiparty elections. In 1994, the DTA won absolute majorities in only two regions (Hardap and Omaheke) and pluralities in only three others (Caprivi, Otjozondjupa, and Kunene), compared with 15 of the 23 electoral districts in 1989. SWAPO's share of the vote rose substantially in Karas (from 30 to 45 percent), Okavango (50 to 80 percent), and Khomas (45 to 60 percent). The UDF failed to hold onto the predominantly Damara district of Kunene, but did score two out of their three victories in the Kunene sub-districts of Sesfontein and Khorikas. Despite the SWAPO percentage advance, and the failures of the minority parties in 1989, it seems clear that most of the changes can be attributed to DTA absenteeism rather than SWAPO winning the votes of former minority party supporters.

The Netherlands

Since 1830 the Parliament of The Netherlands has consisted of two chambers. The First Chamber has never been elected directly and is therefore not considered here. In the earliest period for which one can speak of a constitutional electoral system (1848-1887), the number of members in the Second Chamber was set at one per 45,000 inhabitants. The country was divided into districts and usually two members per district were chosen by limited suffrage. Half of the Chamber was elected every two years, so that in most districts representatives were chosen at each two-year interval. An absolute majority was necessary for election; if no candidate achieved this figure, a relative majority sufficed in the second round.

In 1888 several changes in the electoral system were made. The size of the Second Chamber was set at 100. Single member districts were introduced, although multi-member districts remained in the cities. The last multi-member districts were eliminated in 1897. From that date, only the top two candidates from the first round were allowed to participate in the second round.

One of the two great political questions of the nineteenth and the beginning of the twentieth century in The Netherlands was the struggle for universal suffrage, the other question being the governmental support for religious schools. Both problems were resolved in the so called "Pacification of 1917", as a kind of package deal. The religious political parties obtained a constitutional guarantee of governmental financial support for religious schools. Universal suffrage, one of the main political goals of the Social Democratic Party, first only for males, female suffrage followed in 1919. Proportional representation was introduced, which primarily helped the Liberal parties, who could no longer expect to gain seats under the district system with universal suffrage. Furthermore, compulsory voting was introduced, to insure proportionality.

The electoral system as introduced in 1917 has remained in effect since that time, although some details have been altered. The Netherlands still elects the Second Chamber according to the multi-member candidate list system of proportional representation. Lists of candidates are presented on the ballot. Since 1956 the name of the party or list is placed above the list. The order of the lists is according to the size of the party delegation in the Second Chamber. (For the parties having no parliamentary representation, the ordering is determined by lot). On a list a party may list up to 30 names on the ballot, or twice the number of its incumbent representatives in Parliament, with a maximum of 80. There are 19 electoral districts, but these exist more out of practical reasons and most parties will submit the lists in each of the districts (although the names on the list may vary, and there is no requirement that candidates live in the district or have any relationship with the district).

All votes cast for a candidate on a party list accrue to the total for the party. For the determination of the number of seats to be appointed to a party, the electoral districts play no role; seats are apportioned proportional according to the national vote. On the ballot a black square with a white circle is placed next to each candidate on the list. The voter must fill in the circle next to one of the candidates with a red pencil (for machine ballots there is a lever for each candidate).

In 1956 the number of seats in the Second Chamber was expanded from 100 to 150. The only threshold for obtaining representation in the Second Chamber is the number of valid votes cast, divided by 150, which also determines the electoral quotient (recently about 60,000 votes). Each multiple of the electoral quotient entitles a party to an additional seat.

When each party has received the seats to which it is entitled in this manner, it is generally found that not all seats have been allotted. The seats that remain are distributed by the method of largest average (the so called d'Hondt method). This replaced the largest remainder system in 1933, as the largest average was felt to provide a more precise proportionality. The use of the d'Hondt method does provide an advantage to larger parties. As a partial compensation for smaller parties since 1973 it has been possible to combine lists, both within and across districts, for the determination of the number of seats received.

Once the total number of seats for each party has been determined, the first name on the list is declared elected. The procedure next moves to the second name on the list, and continues until all seats have been filled. The only exception is that a candidate who receives a quarter of the electoral quotient is declared elected automatically (presuming of course that his or her party is entitled to at least one representative). The lists remain in effect between elections and are used to fill seats that have fallen empty. No by-elections are held.

Roughly the same system as explained above is used for municipal, provincial and European elections. Only the First Chamber is chosen in a different way; its members are chosen by the provincial legislatures.

Though commitment to proportionality is extremely strong in The Netherlands, compulsory voting was abolished in 1970. Further proposals to change the system have come as the result of criticism that there is a growing gap between voters and politicians. Committees were appointed to propose changes and the government eventually proposed a mixed system in which half of the seats would be allocated according to proportional representation and the other half according to a five district system. After considerable criticism that this would disturb proportionality and would not necessarily provide a better relationship with the electorate, the government was forced to withdraw its proposal.

New Zealand: Learning to Live with Proportional Representation

New Zealand used to be regarded as a prime example of a country with an FPTP electoral system. However, after two referendums in the early 1990s, New Zealand adopted a mixed member proportional (MMP) voting system in a unicameral Parliament with 120 members. Until the end of 2004, three general elections had been held using the new system.

Why did New Zealand change its electoral system? What led the country to do something that was extremely unusual for any long-established democracy, especially one with an Anglo-Saxon heritage?

For a start, the FPTP system produced highly distorted results in 1978 and 1981. On both occasions the National Party retained office with an absolute majority of the seats in the House of Representatives despite winning fewer votes throughout the country as a whole than the opposition Labour Party. In addition, both elections saw the country’s then third party, Social Credit, win a sizeable share of the votes for very little return (16 per cent of the votes in 1978 and 21 per cent in 1981 won it only one seat and two seats, respectively, in a Parliament that then had 92 seats). The disquiet engendered by these results led the Labour government elected in mid-1984 to establish a Royal Commission on the Electoral System. Its 1986 report, Towards a Better Democracy, recommended the adoption of a voting system similar to Germany’s. The commission argued strongly that, on the basis of the ten criteria it had established for judging voting systems, MMP was ‘to be preferred to all other systems’.

Neither of New Zealand’s major parties favoured the proposal and the matter might have died had the National Party’s 1990 election manifesto not promised a referendum on the topic. In an initial referendum, held in 1992, nearly 85 per cent of voters opted ‘for a change to the voting system’; 14 months later, the new electoral system was adopted after a second referendum in which 54 per cent favoured MMP (while 46 per cent voted to retain FPTP).

As in Germany, in parliamentary elections in New Zealand the electors have two votes—one for a political party (called the party vote in New Zealand) in a nationwide constituency, and one for a candidate in a single-member district. Whereas representatives for single-member districts (called electorates in New Zealand) are elected by FPTP, the overall share of the seats in Parliament allocated to political parties stems directly from and is in proportion to the number of party votes they receive. If a party wins 25 per cent of the party votes, it will be entitled to (roughly) a quarter of all the seats in the 120-member Parliament, that is, about 30 seats. If a party that is entitled to a total of 30 seats has already won 23 electorate seats, then it will be given another seven seats drawn from the rank-ordered candidates on its party list who have not already been elected in a single-member district. Likewise, if a party entitled to 30 seats has won only 11 single-member district seats, then it will acquire another 19 MPs from its party list. There are two thresholds for MMP in New Zealand. To win a share of the seats in Parliament based on the party votes, a party must either win at least 5 per cent of all the party votes cast in a general election or win at least one single-member district seat. In the 1996 general election, five parties crossed the 5 per cent threshold and one won a single-member district seat but did not clear the 5 per cent threshold. Three years later, five parties again cleared the 5 per cent threshold. Two other parties failed to do so but won single-member district seats, which qualified one of them for an additional four seats in Parliament (it had won 4.3 per cent of the party votes cast in the election). In the 2002 general election, six parties cleared the 5 per cent party vote hurdle, and a seventh party won a single-member district seat that enabled it to bring one other person into Parliament from the party’s list.

These figures point to one major change caused by the introduction of MMP. Established, at least in part, to ensure ‘fairness between political parties’, the new voting system has seen the index of disproportionality plummet from an average of 11 per cent for the 17 FPTP elections held between 1946 and 1993, to an average of 3 per cent for the first three MMP elections. Every FPTP election in New Zealand from 1935 until 1993 saw one of the country’s two larger parties—Labour or National—gain an absolute majority in the House of Representatives. One consequence of MMP has been that, in the three elections to date, no single party has won more than half the seats in Parliament. In 1996, the largest party won 44 out of the 120 seats; in 1999 the largest party won 49 seats; and in 2002 the largest party won 52 seats.

Not surprisingly, then, New Zealand has changed from being a country accustomed to single-party majority governments to being a country governed by coalitions. After the first MMP election, two parties formed a coalition government that commanded a small majority (61 out of 120 seats) in Parliament. Since that coalition disintegrated in August 1998, New Zealand has had minority coalition governments that have had to rely on either formal or informal supporting arrangements (negotiated with other parties or, on occasion, with individual MPs) to ensure that their legislative programmes have been able to win majorities in Parliament. One of the other criteria used by the Royal Commission on the Electoral System was ‘effective government’. The commission noted that electoral systems should ‘allow governments ... to meet their responsibilities. Governments should have the ability to act decisively when that is appropriate’. In this regard it should be stressed that MMP governments in New Zealand have had little trouble governing: all have had their budgets passed without any real difficulty, and none has faced the likelihood of defeat in a parliamentary vote of no confidence. At the same time, New Zealand parliaments have fulfilled another of the royal commission’s criteria by also becoming more effective. Governments can no longer rely on (indeed, they seldom have) majorities on parliamentary committees, and there is a far greater degree of consultation—of give and take—between government and opposition parties in MMP parliaments.

The Royal Commission on the Electoral System also envisaged that under MMP the Parliament would represent the Maori (New Zealand’s indigenous Polynesian minority) and other special-interest groups such as women, Asians and Pacific Islanders more effectively. This has happened. In the last FPTP Parliament, Maori accounted for 7 per cent of the MPs. They now constitute 16 per cent of the members of the legislature. The proportion of female MPs has risen from 21 per cent in 1993 to an average of 29 per cent in the first three MMP parliaments. During the period 1993–2002, the proportion of Pacific Island MPs went up from 1 per cent to 3 per cent, and the number of Asian MPs rose from 0 to 2 per cent.

Discarding a long-established voting system is never an easy process politically, nor is it likely to appeal to entrenched interests or to most incumbent politicians. Leading electoral systems scholars have warned that major electoral reforms should not be undertaken lightly. Nevertheless, there is growing evidence that the parliamentarians of New Zealand and the public alike are learning to live with (if not necessarily love) proportional representation. The reforms adopted in New Zealand in the early 1990s and instituted in 1996 seem likely to last for a considerable time.

Papua New Guinea: Electoral Incentives for Inter-Ethnic Accomodation

The South Pacific country of Papua New Guinea (PNG) has used two different electoral systems—the Alternative Vote (AV) from 1964 to 1975, when it was an Australian territory, and FPTP from 1975 to 2002. It has since reverted to the alternative vote again.

Its experience is interesting for a number of reasons. First, PNG is one of the few developing countries with an unbroken record of continuous competitive elections and numerous peaceful changes of government. Second, the change from one electoral system to another has had a series of unexpected consequences which illustrate the different effects apparently similar electoral systems can have. Third, PNG is one of the few countries to have adopted, abandoned, and then re-adopted a particular electoral system.

Papua New Guinea inherited the AV system from Australia and used it for three elections in 1964, 1968 and 1972. Unlike Australia, however, PNG is a highly ethnically fragmented society, with over 850 separate languages and several thousand competing clan and tribal groups.

Its experience lends support to the claims that AV can promote inter-ethnic accommodation and moderation in deeply divided societies by allowing voters to express not just their first choice of candidate but also their second and later choices. Because of the nature of PNG society, under AV most voters invariably gave their first preference to their own clan or ‘home’ candidate. In many seats, however, this was not enough for any single candidate to gain a majority of votes; they needed the second preferences of other groups as well. In order to gain a majority, candidates had to ‘sell’ themselves as a good ‘second-best’ choice to other clan groups—which meant, in general, someone who would be attentive to the interests of all groups, not just their own. It also meant that those candidates who formed alliances and cooperated with each other would often be more successful than candidates who attempted to win the seat from their own voter base alone. This gave many candidates an incentive to act in an accommodating manner to other clans. The mechanics of the system also ensured that the winning candidate would have the support of an absolute majority of voters. In a substantial number of cases, the winning candidate was not the one who had the biggest ‘bloc’ of supporters but rather the one who could successfully build support across several groups.

Thinking that it would be a simpler system which would have similar effects to AV, Papua New Guinea changed to an FPTP electoral system at independence in 1975. However, the different incentives provided by the new FPTP system led to quite different results from those expected. Because candidates no longer needed an absolute majority of votes cast in order to be successful—just more votes than any other group—the candidate from the largest clan would often win the seat outright. There was no incentive to cooperate with anyone else. Electoral violence increased because it was in some candidates’ interests to stop opponents’ supporters from voting rather than to campaign for their second preferences as they had done under AV. Also, because there were so many clans all trying to win the seat, candidates learned that they could be successful with very limited support.

At the 2002 elections, over half of the MPs in the Parliament were elected with less than 20 per cent of the vote. Several candidates who won seats gained only 5 per cent. In an electoral cycle increasingly dominated by concerns about corruption, power and money politics, this led to a range of negative campaign tactics, such as encouraging rival candidates to stand in order to ‘split’ a dominant clan’s voter base. This increased pressure for the reintroduction of AV. In 2003, the PNG Parliament re-adopted what it called ‘limited’ preferential voting for all future elections. Voters will be required to mark a minimum of three preferences.

The Papua New Guinea case illustrates just how dependent much of the accepted wisdom regarding electoral systems is on the structure of the society concerned. Despite having an FPTP electoral system, PNG had a very fluid party system, based on individuals rather than ideologies. All governments under FPTP were weak coalitions, which sometimes changed on the floor of the Parliament as well as at elections. The single-member system of representation resulted in high levels of turnover of politicians from one election to the next, as members could not both be in Port Moresby at sessions of the Parliament and be continually visible in their districts.

Accordingly, a strong sense of accountability on the part of many local members to their electorate developed: without it their chances of re-election would be slim. This matches a strong sense on the part of the electorate that the function of their member is to deliver direct benefits to the community, building on Melanesian tradition that a ‘big man’ ensures that his community shares in his wealth and good fortune. As one member has memorably put it, ‘When people elect me to Parliament, they think I own the Bank of Papua New Guinea.’

Under the AV system, this sense of accountability tended to be spread across a number of groups, thus helping to manage inter-ethnic conflicts. However, this was itself a reflection of the extreme fragmentation of the country’s society.

Poland: Between Fragmentation and Polarisation

The June 1989 elections in Poland played pivotal role in the collapse of communist regimes across Central and Eastern Europe. These elections came about as a result of negotiations between Polish Communists and the Solidarity-led opposition, finalised by the so-called Roundtable Accord in April 1989. The Accord set up a unique (and designed for a singular use only) system of "compartmentalised" elections, with 65 percent of the 460 seats in the Sejm allocated in advance to the Communists and their allies, and the remaining 35 percent subjected to an open contest. In addition, the upper chamber of parliament, the Senate, was re-established, with 100 seats to be filled in a free, unrestricted election - the first fully free and fair election in Eastern Europe in more than 40 years. In the elections, held on June 4 and June 18, 1989, Solidarity won all but one of the seats it contested. The communists acquired their allocated seats in the Sejm, but none in the Senate. The outcome of this election reflected the high level of political polarisation in Poland throughout the 1980's.

The Sejm

The parliament, elected in 1989 to a four-year term, as early as 1990 became obsolete. Being still dominated by the people of the old regime, with apparently slim chances for reelection, it did not go without hesitation, but eventually the date of new elections was set for October 27, 1991. After a long debate, the new electoral law for the elections to the Sejm was adopted by that House on June 28, 1991. Participants of this debate had their overt and covert agendas, and often changed their positions. In general, the post-communist majority in the Sejm, together with some post-Solidarity groupings was in favour of proportional representation, while others (President Walesa and the major post-Solidarity parties) advocated various mixed electoral systems. The rule of the thumb was the stronger the party and the more popular its leaders, the less likely it was to opt for a pure Proportional Representation (PR). Eventually, the considerations stemming from the uncertainty of the election outcome prevailed in deputies' minds, and for the Sejm election a PR system was adopted.

In the system adopted in 1991, the country was divided to 37 districts, with from seven to 17 seats in each, determined according to population. To register its list in a district, a party had to collect at least 5,000 signatures of voters resident in the given district. Seats in the districts were allocated to the parties through the Hare-Niemeyer method, with no threshold. Out of the total 460 seats, 69 were awarded (through the modified Sainte-Lagu method) on a nation-wide base to each of those parties, which: (a) registered a national list (for which a party had to register its list in at least five districts by collecting 5,000 voters' signatures in each), and (b) surpassed the threshold of five percent nationally and/or managed to have their candidates elected in at least five districts. These thresholds were waived as far as the representation of ethnic minorities was concerned. Voters in each district were to show their preference by selecting one party list and placing a check beside the name of the candidate for whom they wished to vote.

The political dichotomy of the 1989 election (Solidarity vs. the old regime), in 1991 was replaced by a highly fragmented polity, with more than 100 parties and quasi-parties contesting both the Senate and the Sejm elections. The elections resulted in a highly fragmented parliament, with the strongest party in the Sejm controlling a mere 13.5 percent of the seats, and no majority coalition of fewer than five parties possible. Altogether, 18 parties and groupings mustered at least two seats each, and additional 11 one seat each. The actual number of actors in the Sejm was in fact lower, due to mergers and coalitions. This parliament in two years managed to generate three Prime Ministers, two governments, and its own early departure, after accepting a motion of non-confidence in Hanna Suchocka government in May 1993. President Walesa, having a choice between dismissing the government or dissolving the parliament chose the latter. Among the last bills approved by the departing Sejm were amendments to the electoral law.

The amendments adopted by the outgoing parliament were designed to limit fragmentation by eliminating weaker parties from the Sejm. Specifically, three devices were put in place: (1) a threshold of five percent for parties and eight percent for coalitions, nation-wide (also the threshold for national list was raised from five percent to seven percent); (2) an increase in the number of districts, from 37 to 52, expanding district magnitude to three to 17 seats; (3) implementation of the D'Hondt formula, advantageous to stronger parties, for allocation of all seats. Support for these amendments among the parties correlated in an obvious (positive) manner with their strength in the outgoing Sejm. For some of the parties that supported the amendments this decision was self-destructive, as only one coalition and five parties surpassed in the Sejm elections the prescribed thresholds.

Thus the electoral reform gave the expected results: the stronger parties were awarded additional seats, the weakest were altogether eliminated from the Sejm. While the fragmentation of the parliament has been overcome, it has been achieved on the expense of serious distortions of proportionality. The ruling leftist coalition that emerged after the election received together only 36 percent of the votes but commanded a majority of 66 percent of the seats. With 34 percent of the votes "wasted" for the parties not represented in the Sejm (mostly from the right side of the political spectrum), the parliament elected in 1993 was perceived as not fully legitimate: it represented (since the turnout was at the 52.1 percent level) only 34 percent of the eligible voters. Thus one potential (and between 1991 and 1993 actual) dysfunction of the democratic order, unstable governments generated by a fragmented parliament, has been replaced in 1993 by another one: a stable government without sufficient legitimacy.

This potential deficiency has been removed in the wake of the most recent, September 21, 1997 elections, conducted according to the same as in 1993 rules. While only five parties and coalitions cleared the thresholds this time, they represented 87.8 percent of all valid votes (with the turnout of 47.93 percent). More importantly, the political balance returned to the Sejm, since several parties of the fragmented right united this time under the Electoral Action Solidarity (AWS) umbrella and won plurality of votes, to form a right-of-centre government with the liberal-democratic Freedom Union. However, the competition of the two major coalitions, the post-communist SLD on the left, and the AWS on the right, led to the weakened support for the centrist parties, and to the re-polarisation of the polity along ideological lines.

The Senate

Since 1989, the voting districts in the Senate elections have been based on geography, with two senators elected in each of 47 out of the 49 Poland's provinces. Only the two most populous provinces have been given three senatorial seats each. Senators are elected by a candidate-centred bloc vote, see Block Vote. In 1989, a majority was needed to win a seat (with a run-off two weeks after the first round). For the subsequent (1991, 1993, 1997) Senate elections the run-off round was abolished (plurality vote replaced majority vote). Interestingly, the level of fragmentation, and - with only minor deviations - the political composition of the Senate have been since 1991 parallel to the party composition of the Sejm.

Presidential and Local Elections

Since 1990, President of the Republic of Poland is elected by a popular vote. If in the first round none of the candidates gains majority, a run-off among two top candidates takes place two weeks later. In the 1995 election, which featured a tight race between the incumbent Lech Walesa and the eventual winner Aleksander Kwasniewski, this arrangement contributed to the re-polarisation of the political spectrum.

In local elections two systems are used in a parallel way: First Past The Post (FPTP) in rural communities and in cities below 40,000 inhabitants, and a party-list PR in cities with more than 40,000 inhabitants.

Palestine: Political Realities Shape the System

The Declaration of Principles or Oslo Agreement, reached in late 1993 between Israel and the Palestine Liberation Organization (PLO), contained a provision for an elected Palestinian Council to be established. The implementation of the Oslo Agreement required the negotiation of a further detailed agreement, the Interim Agreement. This was completed in Taba in September 1995 and included detailed provisions for holding elections to the Palestinian Legislative Council and, separately, for the head of its Executive Authority. The president (Raees) of the Palestinian Authority and the Palestinian Legislative Council were then elected on 20 January 1996.

Preparations for the elections began in 1994 in parallel with the negotiations for the Interim Agreement. The election law and the conduct of the elections were entirely the responsibility of the Palestinians, although some details of the election arrangements were required to be consistent with the provisions of the Interim Agreement. The final version of the law and the major regulations were put in place only in late 1995.

The political context of the election strongly influenced the available options for the electoral system. There was little doubt in anyone’s mind that Yasser Arafat would be elected president, and for the presidential election a single-round FPTP system was adopted with little discussion. The assumption was borne out in practice when Arafat received over 80 per cent of the vote against one other candidate.

The choice of system for the Legislative Council elections was much less straightforward. First, agreement within the Palestinian community on accepting and participating in the Interim Agreement process was not unanimous. The emerging Palestinian Authority conducted lengthy discussions backstage with members of Hamas and other Islamic movements which included the question of their participation in elections. Second, the political party system was embryonic. Fatah had the character of a national liberation movement, a political form for which a continuing need was perceived because of the need for unity in moving into ‘final status’ negotiations with Israel (which were not successful). Some other small parties had formed, but many potential candidates were considering standing independently of Fatah. Third, there were some precedents to hand: local elections had been held in Gaza in the 1940s, using Egyptian procedures, and in West Bank cities and towns in the 1970s, using Jordanian procedures inherited from traditions under the British Mandate. There was pressure in particular to follow Jordanian practice.

The choice of a candidate-based electoral system therefore emerged in response to three pressures: the wish to provide a channel for informal candidacies of persons linked to movements which formally rejected the process; the desire of a number of prominent figures to stand as independents; and the recollection of historic elections. The importance placed on simplicity, transparency, speed of counting and confidence in the results also led to a decision in favour of counting at the polling station, thus eliminating preferential systems such as the Alternative Vote (AV) or the Single Transferable Vote (STV) as options. The perception of where natural boundaries existed on the ground thus led to the choice of the Block Vote (BV), with districts which varied in magnitude from 12 in Gaza City down to one in the small towns of Jericho, Salfit and Tubas.

A further discussion centred on the representation of minorities, in particular the Christian community (which accounted for some 10 per cent of the electorate) and the Samaritans (a concentrated community of a few hundred people near Nablus). Six reserved seats were created within the Block Vote system for Christians in the four districts with the highest concentration of Christians (two each in Bethlehem and Jerusalem, and one each in Ramallah and Gaza City) and one reserved seat was created for Samaritans in Nablus. Christian candidates had the option to declare themselves as Christian. If the Block Vote count showed that there were not sufficient declared Christian candidates among those in the top positions, the candidate with the lowest vote of those who would otherwise have been elected would be replaced by the declared Christian candidate with the next—highest vote—as indeed happened in all four districts. This meant that there were representatives on the Legislative Council elected with fewer votes than some other candidates who were not elected. While there was some debate on this, it was accepted as legitimate in the context of wide representation and in the aftermath of a successful election.

In practice, the BV electoral system achieved much of what was expected of it. Eighty-seven candidates were nominated in Gaza City, but voters coped well with a ballot paper about a metre long. While few candidates associated with those who rejected the peace process stood, at least one member was elected who might be considered as a bridge to those movements. Candidates on Fatah slates gained an advantage, but voters made clear distinctions between more and less popular individuals. Leading independent figures were elected, as were representatives from minorities. Small towns with a fiercely independent identity gained their own representative. The president and the Legislative Council took office in 1996 with a wide degree of legitimacy within the Palestinian community.

Russia - An Evolving Parallel System

The legislative electoral system, which was first decreed by President Boris Yeltsin in September/October 1993, along with the presidential election system, were included within the first post-Soviet Russian constitution, which was narrowly ratified by the voters in December 1993. The Federal Assembly, the legislature of the Russian governmental system, is bicameral. The Duma (the popular assembly) is elected every four years. The Federation Council (the Upper House) consists of one executive and one legislative representative chosen from each of the 89 regions of Russia according to the laws of each region.

The Russian electoral system can be characterized as a classic example of a parallel electoral system, see Parallel Systems. Both party-list Proportional Representation (PR) and First Past the Post (FPTP) voting are used for choosing deputies in the Duma, but there is no adjustment of the party-list representatives to reflect disparities in the overall seat-vote share, as there is in Germany and New Zealand, see Germany: The Original Mixed Member Proportional System and New Zealand: A Westminster Democracy Switches to PR. The total number of deputies is 450, with exactly half selected by PR and half chosen in single-member plurality constituencies. The PR system operates in effect as one constituency, since the votes for political parties are tallied across the entire country. Nevertheless, parties compete regionally on closed lists, in accord with the June 1995 law adopted by the Federal Assembly. A nominee for a national party list of 12 members may also seek election from an FPTP single-member district in the region. Consequently, this can result in another seat for a political party, which wins on the PR ballot. Upon achieving the threshold of at least five percent of the PR votes, seats are distributed according to the largest remainder formula, see The Threshold. In theory, this is supposed to benefit smaller parties, but it does not appear to have had that effect in Russia.

In the 1995 parliamentary elections, only four political parties crossed the five percent threshold, which would make them eligible to be allocated seats from the PR lists. These parties garnered only 50.5 percent of the popular vote and received double the number of seats which would have been distributed had it been a strictly proportional system. Women of Russia, one of the 18 parties which failed to gain party list seats, was a slim 2.3 percent lower in votes than the Yobloco Party, which obtained, by contrast, 31 party list seats. Anomalies also occurred in the single-member constituencies, some of which were won with percentage votes as low as 20 percent when several of the 43 parties competed. Consequently the proportion of wasted votes was very high in the 1995 parliamentary elections.

The development of Russia's new electoral system was characterized by compromises among parliamentarians, the Russian president, and the legacy of past practice. At first Boris Yeltsin decreed that one-third of the Duma would be elected by party-list PR, and the remainder elected from single-member districts as in the former Soviet Union. However, a number of pro-democracy groups in the previous parliament favoured List PR, seeing an advantage for their mostly Moscow-based organizations. After apparently being persuaded that well-organized communist parties would benefit from single-member districts, Yeltsin adopted an evenly-split plurality-PR system in October 1993. At the same time there was substantial agreement on the method of electing the President and the Federation Council, but in 1995 the election of Federation Council members was decentralised so that elections would be held according to each region's electoral laws.

The five percent threshold, intended to inhibit the proliferation of parties, has not worked in Russia and has led to gross disproportionality in the second Duma, elected in 1995. A number of groups have suggested the complete removal of the threshold, as in Iceland, or a smaller minimum percentage, such as the 0.67 percent threshold in the Netherlands, or the four percent in Sweden. Another change would be a move to a fully compensatory MMP system, as is used in Germany. The seats distributed to parties would then reflect the people's PR vote within each region, thus enhancing overall proportionality and strengthening the political party system as a whole.

Candidates for the presidency in 1991 were required to obtain 100,000 signatures, with only seven percent from the same region, for nomination. In 1995, this number was increased to one million signatures. The presidential system specifies that if no candidate wins an absolute majority in the first election round, a second is held between the two leading contenders, and the winner is required to win 50 percent or more votes for election. Four years is the term of office, and there is a two-term limitation. Presidential elections are held in different years from parliamentary elections. One problem with the two-round absolute majority presidential election procedure is that it discourages the formation of party coalitions, unlike one-round plurality elections in which parties tend to form in two blocs, see Electing a President - Two Round-Systems. Holding the presidential election at the same time as the Duma's would further reduce party-splintering and ensure greater accountability by the president and Duma.

Senegal:A Parallel System in Africa

Senegal is one of only a handful of countries in Africa that have undergone a genuine democratic transfer of power as a result of the defeat of a sitting president. At the presidential level the electoral system is a two-round majority run-off system very similar to that used for elections for the president of France. Parties have incentives to put up candidates in the first round, and voters can freely vote for the candidate they really prefer while saving their ‘strategic’ vote for the second round. In 2000 the opposition parties denied President Abdou Diouf a first-round victory and, by previous agreement, united behind the leading opposition candidate, Abdoulaye Wade, to defeat the long-reigning Socialist Party (Parti Socialiste, PS) leader in the second round.

For the legislature, the electoral system was changed from a pure PR system by national list in 1978 to a mixed, Parallel system since 1983. Since then it has been modified numerous times. Most of the changes have been designed to provide for democratic legitimacy by ensuring that the system remains open to some opposition representation while maintaining the ruling party’s majority of seats. Like many mixed systems, Senegal’s relies on a national list for a part of the seats. Unlike most other Parallel systems, the plurality seats, rather than being decided in single-member electoral districts, are decided on the basis of Party Block Vote (PBV) in mostly multi-member electoral districts.

The PR seats (roughly half of the total) are allocated from the votes cast for the national list of each party, using a Largest Remainder Method with the Hare formula. The other seats are allocated by plurality vote in multi-member electoral districts in the 30 departments of the country, with between one and five seats each. The smaller parties and the opposition have always argued for a greater number of seats to be allocated from the national list, while the ruling party has always favoured a balance—ensuring that its domination of the plurality seats plus a proportion of the Parallel seats would enable it to retain power. For the 1998 election the ruling PS once again altered the distribution of seats, adding 20 new plurality seats. The PS won 18 of these and was easily able to maintain control of the legislature, despite the fact that it had only won a bare majority of the vote nationally (50.3 per cent). President Wade, when he was leader of the opposition, argued for greater proportionality in the system and less reliance on the PBV seats, which heavily favour the party in power. For the 2001 elections, Wade, who had bitterly attacked the machinations by the PS, was in a position to alter this inequitable formula. Once in control of the presidency and with power to modify the electoral system, Wade could install a system designed either to be more representative of the voters’ wishes or to maximize the opportunities for his coalition (the SOPI Coalition, led by the Democratic Party of Senegal (PDS)). Coupled with the greater resources now available to his party, including the full weight of the presidency, this made the highly inequitable system he had previously attacked seem suddenly attractive.

The PS and the Alliance of Forces of Progress (AFP), the only other large parties, also felt that they could profit from an emphasis on the plurality PBV side of the seat distribution by potentially winning a plurality in several districts. They too chose to argue for increased weight for the plurality side of the election. The smaller parties pushed for pure proportionality based on a national list system or some compromise that would provide greater opportunities for a better seat-to-vote distribution. The ruling PDS opted for reducing the size of the National Assembly from 140 to 120 and moving from a 70 : 70 plurality : proportional distribution to 65 plurality and 55 national list PR seats. The PDS calculated that as the new party in power it could win a plurality in many departments, thus increasing its share of seats relative to its voter support.

These calculations proved correct. Although the SOPI coalition received just under half of the votes (49.6 per cent), it won 89 of the 120 seats (74.2 per cent) in 2001. The former ruling party, the PS, finished second in terms of share of the vote, with 17.4 per cent, but garnered only ten seats, all in the proportional national list. In third place in the voting, the AFP of Moustapha Niasse (with 16.1 per cent of the vote), passed the PS in number of seats with 11, two of which it won on the plurality side by finishing first in one department. The Union for Democratic Renewal (URD), with 3.7 per cent of the vote, garnered three seats, one of which was a plurality seat in a single-member district in the small department which is the home of its leader. The African Party for Democracy and Socialism/Jef (AJ/PADS), with just over 4 per cent of the vote, won only two seats, both on the national list. Five additional parties were given one seat each on the national list by virtue of having the largest remainders, even though they did not achieve a full quota in votes. The remaining 15 parties which presented lists were excluded from the seat allocation.

The disproportionality in the 2001 election greatly exceeded even the high rates Senegal had experienced under PS rule. In the 1993 elections, 70 seats were allocated by the proportional formula on a national list and 50 in department-level districts using PBV; in 1998, seats were allocated 70 : 70 between the two electoral formulae, and disproportionality rose slightly. It rose sharply in the 2001 elections, the results of which were less proportional than the results of most elections in FPTP systems. As a consequence the legitimacy of the legislature is compromised. The SOPI coalition, which came to power in 2001 on the basis of arguments for democratic reform, has taken a major step backwards with its manipulation of the electoral system to its own advantage.

In Senegal, the objective of the then hegemonic ruling party in opting for a Parallel system was to ensure fragmentation of the opposition by discouraging coordination among parties, minimizing strategic voting and thereby providing an advantage to the largest party. The party furthered this objective by employing a ballot that offers a single choice that covers both the proportional and the plurality vote. Since the allocation of the proportional seats depends on the total number of votes a party or coalition receives, Senegal’s opposition parties have an incentive to present candidates in as many plurality districts as possible. The fact that all votes cast at the district level are added together for the allocation of the proportional seats reduces any advantages of coordination between parties across constituencies, and provides incentives for sincere rather than strategic voting. Thus, the ruling party was able to ensure wins for itself in most plurality districts.

A Parallel mixed system is also used at the local (rural council) and municipal levels. To ensure an overwhelming majority for the winning party on every council, half of the seats are allocated in a single bloc for the whole municipality using the Party Block Vote. The other half of the seats are allocated by List PR using the whole rural community or municipality as one district.

Senegal illustrates the way in which short-term political advantage can be the overriding factor in debates about electoral system change. An opposition that clamoured for change became a government which defended a status quo which suddenly looked beneficial. The potential down side of this is that, if the electoral pendulum swings back to the PS, the heavier emphasis on the PBV part of the Parallel system is almost certain to ensure that this will be reflected in a more than proportional gain of seats for the opposition, leaving the forces in the SOPI coalition back where they were before 2001. Alternatively, this approach can be seen as moving towards a two-large-party system, in which the two beneficiaries ensure that any potential third political force has a huge task to break in.

Thailand: Combating Corruption through Electoral Reform

In 1997 Thailand adopted a new constitution which brought about sweeping changes to its political and electoral landscape. Reforms included the creation of an autonomous Electoral Commission to oversee and administer elections, new rules governing the relationship between the members of Parliament and the Cabinet, and the creation of an elected Senate—the first ever in Thailand. The constitution also replaced the Block Vote (BV) electoral system that had been in place for most of Thailand’s electoral history with a Parallel system made up of FPTP and List PR elements.

Prior to the 1997 reforms Thailand used the BV system to elect the House of Representatives. The Senate was entirely appointed. The country’s electoral districts were broken down into one-, two- and three-seat districts, with most districts having more than one seat. Seats were allocated by province in proportion to population. Voters cast their votes for candidates rather than parties, and were allowed to vote for as many candidates as there were seats in a district. They could not cast all their votes for a single candidate but could split their votes between candidates from different parties. They could also partially abstain by not casting all their available votes. Parties were required to field a full team of candidates for any district they wished to contest (for example, three candidates in a three-seat district). Seats were awarded to the one, two or three candidates who got the most votes on the basis of the plurality rule.

The BV system in Thailand had at least two major implications for the party system. These multi-seat districts had tended to produce multiple parties in each district, which in turn had contributed to the presence of a large number of parties in the House. The average effective number of national parties between 1975 and 1996 was more than six. Not surprisingly, no party ever commanded a majority, making large, multiparty coalition governments necessary. These coalition governments were generally indecisive and short-lived. Reformers hoped that by changing the electoral system they could bring about a reduction in the number of parties and a reduction in government inaction and instability.

Second, the system pitted candidates from the same party against one another in the same district. Although each party nominated a team of candidates, they often tended to campaign against each other rather than trying to get voters to support all of the party team with all of their votes. This intra-party competition undermined the value of party labels to candidates and voters and contributed to making the parties factionalized and incohesive. One reflection of this was the rampant party-switching prior to every election, with attendant allegations of money politics. Intra-party competition, the weakness of party labels and the relatively small districts also encouraged politicians to cultivate and respond to relatively narrow constituencies. During election campaigns vote-buying helped candidates build personal constituencies. In office politicians focused on providing ‘pork’ and particularistic goods and services to their constituencies, often to the neglect of broader policy concerns and thus to the coherence and consistency of government policy. The drafters of the 1997 constitution hoped that through electoral reform they could encourage the development of party cohesion and meaningful party labels, and bolster the incentives of candidates and politicians to respond to broad, national constituencies.

In 1996 the House of Representatives, responding to long-simmering demands from within civil society for political reform, organized a Constitutional Drafting Assembly (CDA). A year later, after a widespread popular consultation and in the midst of a severe economic crisis that quickly escalated into a political crisis, the CDA submitted and the House approved a new constitution. The cornerstones of this new constitution were an elected Senate and an overhauled system for electing the House of Representatives. Gone is the Block Vote system for the House. Following a growing trend, the drafters of the constitution established a Parallel electoral system in Thailand. Four hundred single-member districts replaced Thailand’s multi-member districts. In these districts voters cast a single vote for their preferred candidate. The 1997 constitution also created a second tier of 100 seats elected from a single nationwide district by PR. A party must reach a threshold of at least 5 per cent of the party list votes in order to be eligible for seats in this tier. Each party is required to submit a list of candidates for voters to consider, and voters cast two votes, one for a district representative and one for a party list. Candidates must choose between running in a district and running on the party list. The two tiers are not linked: a party’s seats in one tier are not in any way dependent on the number of seats it has in the other tier.

The 1997 constitution also provided for an elected Senate, the first in Thailand’s history. Two hundred senators are elected using the SNTV system. The electoral districts range from one to 18 seats in size. The Thai version of SNTV also has an added twist. Constitutional reformers wanted to create a Senate that would remain above the messy partisan fray. As a result, senators are constitutionally prohibited from belonging to a political party and are not allowed to campaign for election.

What were the results of these constitutional reforms? As discussed above, one of the drafters’ chief goals was to reduce the number of parties in Thailand—hence the move to single-member districts and the 5 per cent electoral threshold in the party list tier. It appears that this goal has largely been achieved. In the 2001 election for the House of Representatives, the effective number of parties in the legislature fell dramatically from an average of 6.2 before 1997 to 3.1, reflecting both a decline in the number of parties contesting each single-member district and better coordination of parties between districts. For the first time since 1957 a single party, the newly formed Thai Rak Thai party, nearly captured a majority of the seats. It later gained a majority after a smaller party disbanded and joined its ranks.

The drafters also hoped that adding a national party list tier and doing away with intra-party competition would encourage voters and candidates to focus more on party policy positions regarding national issues. This in fact began to occur in the 2001 election. For the first time in recent Thai electoral history, political parties, led chiefly by the Thai Rak Thai party, put significant effort into developing coordinated party-centred electoral strategies. Parties began to differentiate themselves in terms of their policy platforms and in some cases made those differences an important campaign issue.

However, there are reasons to be somewhat cautious in assessing the emerging changes in the Thai party system. First, the shift towards party-centred strategies was primarily confined to the campaign for party list seats, while contests in the 400 single-member districts generally remained candidate-centred affairs. This is certainly no surprise given the electoral system: single-member districts still generate incentives to cultivate personal support networks (although it appears somewhat less than under the Block Vote system). Second, the new electoral system has brought about a dramatic reduction in the average number of votes needed to win a seat. This is the combined effect of adding more seats to the legislature and switching from Block Vote to single-member districts. This weakens the incentives to abandon personal strategies: the smaller the number of votes required to be elected, the more likely it is that individual candidates will employ personal strategies. Finally, the presence of a non-partisan Senate, elected by SNTV, undermines somewhat the attempt to create a more party-oriented electorate.

Obviously, any assessment of the consequences of the 1997 reforms must still be tempered. With only limited data available, it is not possible to determine whether the outcomes of the 2001 and 2005 elections represent new trends or are a reflection of the ‘one-off ’ personality of Prime Minister Thaksin Shinawatra, the leader of Thai Rak Thai. Nonetheless the 2001 and 2005 House elections already mark Thailand as an interesting case study of the consequences (sometimes unintended) of electoral system reform.

Ukraine - The Perils of Majoritarianism in a New Democracy

Ukraine's first elections as an independent state were held under a majoritarian Two-Round System (TRS), see Two-Round System. The collapse of the Soviet system in the early 1990s had precipitated the creation of numerous new parties seeking to establish their democratic credentials and lead Ukraine through economic and political reform. But geographical variations in Ukraine's economic and ethnic structure, coupled with a history of territorial division, led to the formation of different parties in different parts of the country. The result was a situation in which there were many small parties with support bases defined either along regional and ethnic lines, or along those of economic wealth. When Ukraine held parliamentary elections in 1994, most of the parties were ill organized and had only a vague idea of how many supporters they had. One reason for this is that, although Ukrainians on the whole valued democratic politics, there was also a strong popular aversion to organized political activism, given the country's experience of one-party rule under communism.

The weakness of the parties at the outset of multi-party competition meant that electoral institutions were especially important in shaping the young party system. According to the electoral law that governed the 1994 contest, one deputy was elected from each constituency, and a run-off was held between the two candidates who received the highest number of votes if no candidate gained an absolute majority in the first round. Many commentators at the time saw the Two-Round System as an ideal means of limiting the number of parties in parliament while at the same time giving small parties a greater chance of being elected than they would have under a First Past the Post system (FPTP). Another perceived advantage of the system was that it would encourage the formation of tactical agreements between like-minded parties in the second round, to maximize the overall representation of their combined interests.

But the results of the elections demonstrated a number of flaws in this reasoning. Firstly, the geographical heterogeneity of partisan support led to the election of many deputies with narrow regional concerns, often associated with the interests of a specific ethnic group or economic sector. When the parliament assembled, it contained deputies from no fewer than 14 parties, a far larger number than that envisaged by the proponents of the majoritarian electoral law. Moreover, the tendency of majoritarian systems to exaggerate the seat share of large parties meant that although the re-vamped Communists gained only 13 percent of the vote, they won 23 percent of the seats and were thus considerably over-represented relative to their true electoral support. This "seat bonus" effect did not operate for the smaller newly formed parties, who mostly received fewer seats than their popular vote may have indicated. Secondly, the elections did little to consolidate the party system; most parties were reluctant to strike second-round deals amongst themselves, because they over-estimated their electoral strength and believed that they would perform best on their own. And thirdly, the preservation of single-member districts allowed many local officials and well-known local figures to win seats without having to associate themselves with an organized party. As a consequence, half of the deputies elected were independents. The large number of parties in parliament and the relatively small proportion of party-affiliated deputies generated a considerable amount of fluidity in the structure of parliamentary factions. This has led to unpredictable outcomes. It has weakened democratic accountability, and it has lowered the parliament's esteem in the eyes of many voters.

A further problem with the Ukrainian electoral law is that it included two stipulations not found in most laws of this type: electoral participation had to exceed 50 percent for the election in a given constituency to be declared valid, and the winning candidate had to receive an absolute majority of the vote. These requirements meant that deputies were not elected at all in about a quarter of the constituencies; low turnout caused many elections to be declared invalid, and in many more cases neither of the candidates in the run-off election won over 50 percent of the vote, since many people voted against both candidates as a form of protest. The process of filling the empty seats carried on for over two years, generating considerable popular disaffection. Moreover, fluctuating numbers in the legislature added to the unpredictability of results, and several regions of the country were left severely under-represented for much of this period.

Following the 1994 elections, there was a general consensus that it would be desirable to move toward a more proportional electoral system so as to reduce the number of independent deputies, stabilise the party system, and promote more predictable legislative behaviour, see PR Systems. The electoral law introduced for 1998 is a semi-proportional parallel system, by which half the deputies will be elected by FPTP in single-member districts, and half from national party lists, with a three percent threshold for representation.

The most important conclusion to be drawn from the Ukrainian case is that, although proportional systems can often cause a proliferation of parliamentary parties in developed democracies, majoritarian laws also allow a large number of parties to enter parliament when parties are weakly entrenched and geographically distinct, which is the case in many new democracies. Furthermore, in Central-Eastern Europe majoritarian systems do little to help consolidate new party systems, because lack of widespread party identification encourages the election of independent candidates who can blur the balance of party strength in parliament and destabilize the legislative process. Finally, majoritarian systems give a distinct advantage to those parties that do have established organizational and support bases, such as those found in the former one-party states of Africa, Eastern Europe, and the former Soviet Union.

United Kingdom: Electoral System Experimentation in Cradle of FPTP

The classical First Past the Post (FPTP), single-member district, electoral system that is so strongly associated with Great Britain did not in fact come into widespread use for Westminster elections until 1884-1885 - a full 50 years after the First Reform Act of 1832, which marked the beginnings of representative democracy in the UK. Up until 1867 most members of the British House of Commons were elected from two-member districts by the Block Vote who served to compound the seat bonuses given to the larger parties. The Second Reform Act of 1867 introduced the Limited Vote (in which electors had one fewer vote than the number of seats to be filled) for the election of 43 members of the Commons, chosen from 13 three-member districts and one four-member seat.

The Third Reform Act of 1884-1885 abolished these Limited Vote seats and FPTP became established as the dominant system. Even today, and despite Westminster's reputation as the birthplace of FPTP, the system is not used throughout the United Kingdom. The Single Transferable Vote form of PR was re-introduced in Northern Ireland, after a 50 year absence, for local government elections in 1973 in an attempt to craft incentives for accommodatory behaviour between the political representatives of the Nationalist and Unionist communities, advantage the moderate and non-sectarian middle, and ensure adequate representation of the minority Catholic community.In the same year STV was used to elect the ill-fated Stormont Assembly - which had been created to give the people of Ulster a degree of self-governing power. Nearly a quarter of a century later, in May 1996, a new body charged with finding solutions to the province's troubles, the Northern Irish Peace Forum, was elected by PR in order to give rise to the most representative body possible. Ninety Forum members were elected from 18 list PR districts of five members in size, while the top 10 parties in terms of votes won across Ulster were awarded two additional seats in the assembly. Since 1979 Northern Ireland's three members of the European parliament have been elected by STV while, at the same time, Britain's 84 English, Scottish, and Welsh MEPs have been elected by FPTP.

The proliferation of different electoral systems in use in the UK has meant that electoral reform, for all tiers of British government, has become an increasingly debated issue. In July 1997 the new Labour government, led by Prime Minister Tony Blair, announced that they would present legislation to change the electoral system for British members of the European parliament to a form of regional list PR in England, Scotland, and Wales, while leaving unaltered the PR STV system in Northern Ireland. Similarly, the proposed Scottish and Welsh assemblies, which will have a degree of autonomous law-making power devolved from the Westminster parliament, are to be elected by PR methods if they are approved by the Scottish and Welsh peoples in September 1997 referenda. Both assemblies are to have Mixed Member Proportional systems which retain FPTP seats based on the current Westminster single-member districts, but include district-based PR lists which will compensate, to some extent, for any overall disproportionality. The proposed Welsh Assembly will have 40 FPTP single-member seats and 20 list PR seats, while the proposed Scottish Assembly will have 73 FPTP seats and 56 list PR seats. No set threshold for representation has been agreed upon but the Welsh Assembly will have an effective threshold of just under five per cent for a party to win a list seat while in Scotland parties will need far fewer votes to gain representation - probably closer to 1.5 per cent of the total vote. Lastly, STV has been proposed by the Fabians (an influential Labour-affiliated policy institute) for local government elections. But it is unlikely that electoral system reform will be seriously considered for local government in this parliament's lifetime - not least because the government's agenda for constitutional reform is already so over-loaded.

However, the overwhelming focus of electoral reform remains the House of Commons and at the time of writing Britain appears closer to changing her FPTP system than at any time since 1917. In that year a proposal to introduce the Alternative Vote (AV) for two-thirds of the parliamentary seats, and the Single Transferable Vote (STV) for the remaining one-third of seats, was narrowly defeated after a stalemate between the House of Lords and House of Commons. A second attempt to move to AV was rejected by parliament in 1931, and it was not until the 1970s that electoral reform muscled its way back on to the British political agenda. In 1976 the Hansard Commission on Electoral Reform, chaired by the former Conservative cabinet minister Lord Blake, recommended that a Mixed Member Proportional (MMP) system be used for parliamentary elections, with three-quarters of the members being elected by FPTP and one-quarter from regional PR lists. The calculation for list seat allocations would take place at the national level and these seats would compensate for any disproportionality in the overall results of the single member district seats.

After four consecutive defeats for the Labour party (1979, 1983, 1987, and 1992) the previously solid Labour support for FPTP began to fracture and in 1990 the leadership set up a commission, chaired by Professor Raymond Plant, to investigate electoral system reform options. The Plant Report (1993) recommended a switch to a sibling of the Alternative Vote which they called the Supplementary Vote - the same system used to elect the Sri Lankan president. While this proposal was never officially adopted by the Labour party in opposition they did nonetheless adopt a policy that, when returned to office, they would hold a national referendum on electoral system change. This policy was given teeth in a joint agreement on constitutional reform between Labour and the Liberal Democrats (who had consistently advocated a switch to a PR) announced on the eve of the 1997 British general election.

The debate over reforming the way members of the House of Commons are elected reflects the First Past the Post versus Proportional Representation debate which has underlain much of the discussion of British constitutional practice throughout this century. The criticisms of the current FPTP electoral system have been restated many times. First, FPTP in the UK has led to some highly disproportional results where minority parties received far fewer seats than their percentage vote might have indicated and has led to situations where the "losing" party, in terms of votes won, became the winning party in term of seats won and thus formed the government.

The Liberal Party, then Liberal-Social Democratic Alliance, then Liberal Democrats, have been the most victimized on the first count - although over the last four elections the disparity between the third party's vote and seat share has been decreasing. In 1983 the Liberal-SDP Alliance won 25.4 percent of the vote but only 3.5 percent of the seats. In 1987 the Alliance won 22.6 percent and 3.4 percent of the seats. In 1992 the newly formed Liberal Democrats won 17.8 percent of the votes and 3.1 percent of the seats, but in 1997, utilising more sophisticated targeting techniques and benefiting from the tide of anti-Conservative feeling, the Lib Dems were able to win 6.5 percent of the seats with 16.7 percent of the popular vote. The uphill struggle that new parties face under FPTP was dramatically illustrated in the 1989 UK European elections when the UK Green Party won 15 percent of the vote but not a single seat. The second anomaly, of one party winning most votes but forming the opposition, has happened twice in the post-war period. In 1951 the Labour Party won more votes but the Conservatives won most seats and formed the government, while in February 1974 the indignity was reversed with Labour forming the government after the Conservatives had polled most votes.

A second powerful criticism leveled at the British FPTP system has been its inability to adequately represent the nation along lines of gender and ethnicity. Up until 1997 fewer than ten percent of British MPs were women, although Labour's vigorous promotion of women parliamentary candidates and their subsequent landslide victory did nearly double the number of women MPs to 18.1 percent in the 1997 parliament. Ethnic minorities in Britain have been similarly under-represented. Most parliaments preceding the 1987 election were all white, and the four Black and Indian-English MPs elected in that year represented less than 0.5 percent of the total. While Black and Asian representation has increased over the last three elections their numbers in parliament remain substantially below their proportion of the UK population as a whole.

Opponents of FPTP have also cited destabilizing swings in economic policy which arose from the alternation of Conservative and Labour governments between 1945-1979, but the Conservatives 18 unbroken years in office (1979-1997) and Labour's drift toward the fiscally moderate centre has tended to weaken this argument. Finally, some PR advocates have disputed the fact that FPTP creates a strong geographical link between elector and representative in the UK, arguing that many safe Conservative and Labour seats are effectively "rotten boroughs" where MPs have little incentive to make themselves accessible, and that the urban centres of the UK are now so totally dominated by Labour MPs that all other party supporters are effectively disenfranchised.

In contrast FPTP in Britain is defended particular because of its single-member districts and encouragement of a "dominant two-party system". Supporters of the status quo find the single constituency member sacrosanct and argue that this relationship of accountability between a voter and their MP is the bedrock of British democracy. Opponents of PR also point to the fact that all, bar one, UK governments in the post-war period have been single party governments and predict that the coalition governments, which would most likely result from a PR system, would be destabilizing to the country as a whole. Related to the previous point is the argument that FPTP provides a barrier against the fragmentation of the party system, which might involve the break up of the major parties (for example, a split in the Conservative Party between "pro-" and "anti-" European wings). Finally, FPTP is praised for denying a platform to extremist parties such as the National Front and British National Party.

Reform of Britain’s FPTP system for parliamentary elections has become increasingly likely with the formation of a Conservative-Liberal-Democrat coalition government in May 2010. A referendum will be put to the British electorate on May 5, 2011, asking voters to decide between FPTP and a new alternative vote system (AV, also known as instant runoff voting).

United States: Ethnic Minorities and Single-Member Districts

Single-member districts (SMDs) are deeply rooted in American political tradition. From the founding of the United States in the eighteenth century to the present, electoral representation has been grounded on the concept of territorial units and subunits. Americans have always thought of popular sovereignty in spatial terms, beginning with the original conception of the U.S. Constitution as a compact among sovereign states and continuing within the states to the valorization of county and municipal government autonomy or "home rule." The Constitution does not specify how popular elections should be structured, and the states have experimented with a variety of single-member-district, multimember-district and at-large forms. But SMDs frequently, if episodically, have been the method of choice for elections at all levels, federal, state, and local, because they enable smaller, geographically situated communities to send their own representatives to larger legislative assemblies. Conversely, multimember districts and at-large elections have been employed when ruling majorities wanted to emphasize the corporate identity of particular jurisdictions and to suppress partisan or ethnic "factionalism." At-large voting rules such as majority-vote requirements, anti-single-shot laws and numbered places were used to maximize the power of ethnic majorities to control all the seats in their legislative bodies.

Historically blacks have been the primary targets of vote-submergence devices in the U.S. The United States is the only modern democracy founded on the institution of slavery, and blacks are entrenched in its Constitution and political institutions as an internal national "other." Slaves were non-persons, and even free blacks were non-citizens. After the Civil War and Reconstruction, blacks in the South were systematically terrorized during elections and, around the turn of the century, disfranchised altogether. The all-white Democratic Party primary became the only election that mattered, and it turned the "solid South" into a region of one-party states. International pressures of the Cold War and the NAACP's litigation campaign against legalized racial segregation eventually succeeded in striking down laws which denied blacks the vote and barred them from primary elections. Thereafter, many majority-white jurisdictions, in and out of the South, resorted to at-large and multimember election schemes to minimize black electoral influence.

The U.S. Supreme Court responded to the post-World War II reexamination of American nationality by elevating the constitutional importance of the individual. In 1963 and 1964 the Court reversed its longstanding refusal to get involved in redistricting controversies and granted relief to white urban voters complaining about the refusal of state legislatures, dominated by underpopulated rural districts, to redistrict themselves. The Supreme Court relied on the Equal Protection Clause of the Fourteenth Amendment to announce the rule of one person, one vote, defining the individual citizen as the basic unit of electoral politics. However, by making the under-weighting of a person's vote justifiable, the Court opened the door to claims that voting strength could be diluted by non-mathematical means, in particular by electoral structures which allowed a bloc-voting white majority to deny a black citizen any opportunity to choose a representative in the state or local legislature. The Supreme Court responded by instructing lower courts to prefer SMDs when they ordered redistricting of malapportioned legislative bodies, and in 1973 it declared unconstitutional Texas' use of multimember legislative districts, specifically because they denied black and Latino voters an equal opportunity to elect candidates of their choice.

All these electoral reforms were wrought by judicial reinterpretation of the Constitution. Meanwhile, in 1965, prodded by the confrontational mass politics of the Civil Rights Movement, Congress passed and President Lyndon Johnson signed the Voting Rights Act, which enabled most blacks in the South to vote for the first time. The conditions that would warrant judicial relief from minority vote dilution became the subject of intense and increasingly complicated litigation, both with respect to at-large or multimember-district elections and with respect to allegedly gerrymandered SMDs. In 1980 the Supreme Court held that racial minorities must prove that a challenged election structure was designed or maintained intentionally to dilute their voting strength. Congress responded with the Voting Rights Act of 1982, which created a statutory entitlement to judicial relief from election structures which had the effect or "result" of diluting the voting strength of protected minorities, defined as racial groups and "persons who are American Indian, Asian American, Alaskan Natives or of Spanish heritage." The 1982 Voting Rights Act, helped along by a 1986 Supreme Court decision which streamlined the proof it required, sparked widespread changes from at-large elections to SMDs, through both litigation and legislation.

By the time the 1990 census rolled around, nearly every state and local redistricting authority was preoccupied with the task of drawing "minority-majority" SMDs that would comply with both the constitutional rule of population equality and the anti-vote dilution mandate of the Voting Rights Act. The new SMDs produced remarkable gains in office holding for both African Americans and Latinos. The number of black elected officials nationwide grew from 300 in 1964 to approximately 8,000 in 1993, although this figure still constituted less than two percent of all elected officials in a country where blacks account for twelve percent of the population. Since passage of the 1965 Voting Rights Act, the number of African-American members of Congress had increased from nine to thirty-eight, and majority-black SMDs were responsible for all seventeen of the African Americans elected to Congress from the eleven Southern states of the old Confederacy. After the 1994 elections, under a new redistricting plan negotiated by black political leaders, Alabama became the first and only Southern State ever to achieve black proportional representation in both houses of its Legislature.

The nationalist backlash provoked by this surge in majority-black and majority-Hispanic SMDs probably was inevitable. The way SMDs are drawn necessarily defines the constituencies that are deemed to be relevant for purposes of representation in legislative assemblies, and it does so in strictly geographic terms. Seldom are redistricting choices politically irrelevant, mere administrative devices for cumulating individual voter preferences. Rather, they declare who the operative national subcommunities shall be and how much power they will enjoy in the lawmaking process. In the United States, counties, municipalities, and recognizable neighborhoods have been the traditional building blocks for redistricting, except when it was expedient to ignore their boundaries for the sake of submerging the electoral influence of African Americans and other ethnic minorities. Now it has become necessary to split up traditional political subdivisions to create districts with African-American or Latino majorities, because in the U.S. people of color have no clearly discernible "homelands." Although they frequently are clustered in ethnically identifiable neighborhoods, these residential enclaves are dispersed among more populous, predominantly white neighborhoods. The result in some cases has been very irregularly shaped, noncompact majority-black or majority-Latino districts which, although they were no more bizarre than some majority-white districts, unmistakably signaled racial or ethnic designs.

The most contorted black and Latino districts quickly drew court challenges from white voters, who contended they violated a radically "colorblind" interpretation of the Constitution. In 1993, the Supreme Court issued the first of a series of decisions which established "an analytically distinct" constitutional cause of action that could be used by individual citizens who wished to challenge "racially gerrymandered" SMDs. Plaintiffs would not have to bear the heavy burden of proving that because of the challenged districts their votes were denied or abridged or that their voting strength was diluted. Instead, the Court recognized a presumptively stigmatic harm ensuing from districts which were drawn for the "predominant" purpose of race and which could not be justified as a "narrowly tailored" effort to serve a "compelling state interest." Such districts are unconstitutional, said the Court, because they presume that all members of the ethnic minority think and vote alike and share the same political interests, a message the Court fears will encourage racial "balkanization" of the electorate. This new gerrymander jurisprudence, which aims to address perceived harms to national unity rather than to the individual plaintiff, has produced court orders striking down several majority-black and majority-Latino SMDs at the Congressional, state, and local levels. The new constitutional districting rules have been created and reaffirmed by the same narrow, five-justice Court majority over the vigorous dissents of four justices, who contend that they offend both substantive justice and the proper limits of judicial review.

Justice Felix Frankfurter warned about the perils of the judiciary entering the "political thicket" in his dissent from the first one-person, one-vote case in 1963. He may be vindicated by the incoherence of the Supreme Court's gerrymandering principles. Surely nothing could be less appropriate for resolution by judges than questions about how the sovereign people should define themselves in a multi-ethnic democratic republic. In its rush to prevent state legislatures from assuming that all African Americans think alike, the Court has yet to confront the converse proposition: What if African-American or Mexican-American or Asian-American or Native-American citizens in a particular state or locale actually do share the same political interests and freely associate to assert them through their elected representatives, through their community institutions or through political organizations - perhaps political parties? To suggest that citizens of color are constitutionally prohibited from negotiating for their own SMDs would contravene historical, constitutionally protected notions of political freedom in the U.S. This is an entirely different question from whether members of an ethnic minority can demand that such districts be created as a matter of legal or constitutional right.

But these are serious questions, which advocates of "majority-minority" SMDs themselves are only now being forced to address. There was never a consensus among them about the political limits or normative endpoint of the voting rights they pressed into remarkably successful service. Today, most advocates of SMDs designed to produce voter majorities of a particular ethnic group defend them as necessary responses to the "unfortunate" reality of ethnic divisions in the national fabric. They share with the opponents of majority-minority districts an underlying commitment to the vision of the United States as an immigrant nation, one in which newcomers and their descendants voluntarily assimilate in the established institutions of public political and social life while retaining the right to preserve their ethnic distinctiveness in strictly private institutional ways. Even private (white) ethnic associations were under pressure to disappear during the "melting pot" era of Anglo ascendancy, which extended at least through World War I. A distinct change in American identity was wrought by World War II, however, when the descendants of other European nationalities placed their stamp of ownership on the American nation and the Anglo-American political traditions they had adopted. The full implications of this national redefinition were largely submerged, as they were throughout the world, in the empires created by competing statist ideologies during the Cold War. That has all changed now, and the U.S. is not immune from the winds of ethnic nationalism that are sweeping the globe.

Today, Americans of German and Irish ancestry outnumber those of English descent. They now sit in the front benches, along with Southern-European and Eastern-European Americans, including secular American Jews, where together they have become the most passionate defenders of their adopted English language and Anglo-American Constitution. Ethnically identifiable SMDs are an embarrassment to these Americans and a threat to their national vision. The right wing of the immigrant nation supports the current regime of suppressing and delegitimising SMDs that have all too obvious racial or ethnic designs, while the left wing either defends majority-minority districts as temporary integration tools or urges that they be replaced with multimember-district schemes using semi-proportional or single-transferrable-vote rules. The growing number of PR proponents also criticize SMDs because they can make it easier for incumbents to get re-elected, engendering a lack of accountability which hurts ethnic majorities and minorities alike. But PR systems are not invulnerable to the same charges often leveled at majority-minority SMDs, that they encourage ethnic polarization and threaten destabilization.

Left out of this immigrant debate and its common objectives of national uniformity, however, are Americans of color, especially the descendants of African slaves. Some scholars now acknowledge that white supremacy has always been a more powerful defining characteristic of American citizenship than any of the more openly debated versions of liberal pluralism and civic republicanism. Many white Americans are simply disturbed or even frightened by black control of the political units in which they reside. For African Americans, a more inclusive immigrant nation may be neither realistic nor an acceptable remedy for centuries of caste exclusion. They may favour renegotiations of American nationhood on terms that at last acknowledge their distinctiveness and accord them full dignity and free agency. The periodic redrawing of SMDs may be one of the best ways of forcing their national demands onto the table, which could explain why a hostile Supreme Court majority has constitutionalized the issue in hopes of squelching the debate. Proportional representation systems may afford African Americans equal participation in legislative bodies, but by sidestepping the constitutive inter-ethnic dialogue redrawing SMDs requires they may actually impede the historical quest of descendants of slaves for complete freedom. PR proposals by some members of the Congressional Black Caucus have not resonated strongly in the black community. On the other hand, it is easy to imagine how the descendants of conquered indigenous peoples and of non-white immigrants might have entirely different views of which election structures best suit their personal and collective agendas in an increasingly diverse U.S.

A just resolution of these conflicting, often incommensurable ethnic positions on electoral structures and their underlying national visions can be achieved only through mutual consent to compromises, which must be incomplete and provisional so long as we value the liberal ideal of individual freedom to shape and reshape one's own cultural and political identity. The negotiations required to reach agreement on such formative questions are particularly difficult to start and to sustain in the United States, because for so many Americans their national identity is invested in a sacred, written Constitution, which for all practical purposes can only be reinterpreted, not renegotiated. Not surprisingly, the greatest progress toward national consensus usually has been achieved through democratically negotiated compromises outside the constitutional context, as with the Voting Rights Act, for example. Now, with considerable encouragement from "colorblind" conservatives, some members of the Supreme Court are suggesting that what they consider to be overzealous implementation by the democratic branches of federal and state governments may call into question the constitutional validity of the Act itself. And the occasion for this constitutional confrontation will be the battle over legislative redistricting. Thus, if the American experience with SMDs as an instrument of political empowerment for ethnic minorities holds any lessons for other democracies, they would include the importance of the particular national context, of respect for its political traditions and the particular situations of subnational groups within them, of the opportunities for gaining the widest possible consensus in making decisions about election structures, and, most of all, of humility when it comes to expectations of lasting solutions.

United States: Total Recall - the election of Schwarzenegger in California

A detailed summary of the election of film star Arnold Schwarzenegger as Governor of California, in the most high-profile recall campaign seen in any democracy. The document gives an overview of the recall mechanism in California and of the campaign to oust Governor Gray Davis.

A high-profile recall case involving a prominent figure has increased public exposure to the recall mechanism. In California, the actor Arnold Schwarzenegger was elected Governor after a successful recall campaign to oust Governor Gray Davis. Gray Davis was re-elected Governor of California in November 2002. However, within less than a year he had been ousted from his post and replaced by a man previously more famous for his acting and his muscles. So how did it come to be hasta la vista for Governor Davis in such a short space of time?

The recall mechanism in California

History

Like many other US states, Californian democracy provides for the use of the recall mechanism. Adopted into the Californian constitution in 1911, the recall mechanism is a process by which the electorate can seek the removal of elected public officials before the end of their terms of office. Prior to 2003, the recall mechanism had been used in California on numerous occasions. Several local government officials have been recalled, and four state legislators have been recalled, in 1913, 1914 and 1995 (twice). However, whilst previous state Governors have faced some level of recall attempt in the last 30 years, Governor Davis was the first Governor to face a recall election.

Launching a recall initiative

To launch a recall initiative, proponents are required to file a notice of intention with the California Secretary of State's Office. The notice must include a statement explaining in no more than 200 words why the proponents are seeking a recall, and must give the names, signatures and addresses of either a minimum of 10 people or the number of people required to have signed the nomination of the officer who is subject to the recall, whichever is the greatest. Proponents of the recall must be registered voters within the electoral jurisdiction of the officer they seek to recall. Within a week of the notice being filed, the officer in question is given the opportunity to make a 200 word statement in response to the initiative.

Circulating the petition

Once the notice has been certified by the Secretary of State, proponents are required to collect a number of signatures equivalent to 12% of the votes cast for the officer the last time the office was on the ballot. Signatures must be collected within 160 days, and the petition must be circulated in at least five Californian counties. In the 2003 recall, the recall petition was certified for circulation on 25 March 2003, giving proponents until 2 September 2003 to collect 897,158 signatures. Signatures are only deemed to be valid if they are from voters registered in the electoral jurisdiction of the officer who is the subject of the recall; similarly, only such registered voters are qualified to circulate the recall petition.

Reporting and verifying petition signatures

Proponents of the recall are required to file the petition with the election official in each county in which the petition is circulated. Whilst the petition is being circulated, county election officials are required to report to the Secretary of State on the progress of the petition; the first report must be submitted within 30 days of the recall being initiated, and every 30 days thereafter. The report must advise on the number of signatures collected in the most recent 30 day period, the total cumulative number of signatures collected, and the total number of valid signatures collected. County election officials are not required to begin verifying signatures for validity until 10% of the total number of signatures required have been collected. Once the verification process begins, a random sample of signatures must be checked. Where more than 500 signatures are reported at any one time, either 3% of signatures submitted or 500 signatures must be checked, whichever is the smallest.

The recall election

Once the Secretary of State certifies that the petition has qualified, the Lieutenant Governor is required by the California Constitution to set the date of the recall election. The election must be held in the period between 60-80 days after the date on which the Secretary of State certified that the petition qualified, unless there is a state-wide election within 180 days of that date (in which case the Lieutenant Governor has the option to hold the recall election on the same day).

In California, the recall and the vote for a successor to replace the officer should the recall succeed are included on the same ballot. The officer facing recall is not entitled to stand as a candidate for re-election. Therefore, if in the first vote the incumbent officer receives the support of more than 50% of the electorate, the recall is defeated and the second vote is irrelevant. However, if a majority vote to recall the officer, the candidate who receives most votes is elected. At the 2003 recall election, there were also two state wide initiatives on the ballot. This differs from other US states and other countries, where once a petition has qualified, a separate yes/no vote on whether or not to recall the officer will be held before a vote on a successor is held. Critics of the Californian process argue that this causes confusion for voters. In addition, it also raises the possibility that if an officer is recalled, his successor could be elected with a less legitimate democratic mandate: for example, if an officer is narrowly recalled on a vote of 51% in favour, 49% against, but the winning candidate on the second vote receives the support of 37% of voters, the new Governor arguably has the support of 12% fewer voters.

The 2003 recall campaign

The case against Gray

The campaign against Gray Davis was initially launched by an organisation called the People's Advocate, an anti-tax organisation headed by Ted Costa, the official proponent of the recall effort. The main charge against Gray was that he had mismanaged the Californian economy, creating a budget deficit of over USD 30 billion and the need for large tax increases. The efforts of the People's Advocate were supported by a number of Republican Party activists and by other California political parties such as the Libertarian Party and the American Independent Party. In May, the recall campaign was boosted when Republican Congressman Darrel Issa launched his own recall effort, Rescue California, backed by significant funding and the expertise of political strategists. Rescue California was instrumental in collecting signatures for the recall petition, organising its efforts on a professional basis and claiming to have collected around 70% of the valid signatures submitted. Opponents of the recall included labour leaders and public sector employees, and were led by an organisation called Taxpayers Against the Recall. They argued that using the initiative would set an expensive precedent, and that recall should only be used in the case of abuse of office or corruption.

The announcement of the election

The recall petition was certified for petition on 25 March 2003. By the middle of July, recall proponents claimed to have gathered 1.6 million signatures, well in excess of the 897,158 needed to secure the recall ballot. On July 23, Secretary of State Kevin Shelley verified that 1,356,408 of the signatures were valid. The following day, Lieutenant Governor Cruz Bustamante announced that the recall ballot would be held on October 7.

Candidates for the election

One notable feature of the 2003 recall election was the large number of candidates who stood for election. Nomination procedures for a recall election are required by the constitution to be virtually the same as for a normal election to the office of Governor (except that papers have to be filed no less than 59 days before the date of the recall election). Candidates at the 2003 recall were therefore required to collect 65 nomination signatures and pay a USD 3,500 fee (or submit USD 10,000 in lieu of the fee). The low threshold for nominations meant that 135 individuals stood as replacement candidates, thus ensuring an extremely long ballot for voters to consider.

The Republican and Democrat parties

The most prominent of the candidates was undoubtedly Arnold Schwarzenegger, film actor and husband of a member of America's most famous political dynasty, Maria Shriver, niece of assassinated President John F Kennedy and Bobby Kennedy. Following considerable speculation about whether he would join the race to succeed Gray Davis, Schwarzenegger, a Republican, used an appearance on the Tonight Show on 6 August to announce his candidacy. His entry into the race prompted other high-profile Republican candidates to drop out of the race to be Gray's successor. The Democrats faced a more difficult position in choosing whether or not to stand a replacement candidate. On the one hand, their main aim was to win the recall ballot and defeat the recall proposition. It was argued that achieving this result would be helped by the presentation of a united front, with no Democratic candidates for successor. However, an alternative argument was formulated along the lines that, should the recall succeed, it was necessary to have a Democratic candidate on the ballot as a potential successor. The debate was settled when, on August 7, Cruz Bustamante entered the race to offer a prominent Democratic candidacy.

The campaign controls

Different controls applied to the recall element of the campaign, and the campaign to promote replacement candidates. The issue of recall is treated by campaign finance law in California as equivalent to a ballot measure, whilst the question of a successor is treated in accordance with controls on normal candidate elections. In practice, this meant that whilst Governor Davis was not subject to any contribution limits on the amount of money he could accept in trying to defeat the recall measure, replacement candidates were subject to contribution limits of USD 21,200 and an expenditure limit of USD 10,624,000.

The outcome of the recall process

At the election on October 7, the measure to recall the Governor was passed by a vote of 55.4% against a no vote of 44.6%. Of the replacement candidates, Arnold Schwarzenegger achieved 48.6% of the vote (more than had supported Davis), his nearest rival Crus Bustamante achieving 31.5%. Governor Gray Davis was duly recalled, and Governor-elect Schwarzenegger sworn in on 17 November. The total cost of running the recall election, including providing voters with a state voter information guide, was estimated to be just under USD 70 million.

South Africa: Electoral Systems, Conflict Management and Inclusion

The National Assembly parliamentary and provincial elections held in South Africa in 1994 marked the high point of a period of tumultuous change from authoritarian rule to multiparty democracy in Southern Africa as a whole. At midnight on 27 April 1994 perhaps the most despised flag in Africa was lowered, heralding the end of 300 years of colonialism and four decades of apartheid. Those first multiparty democratic elections opened the stage to political movements which had been driven underground by the Pretoria regime’s policy of racial divide and rule. Nelson Mandela’s African National Congress (ANC) was poised on the threshold of power; the Pan-Africanist Congress of Azania (PAC) was challenging it within the same community, while Mangosotho Buthelezi’s Inkatha Freedom Party (IFP) hoped to build on its hegemony in the north of the province of KwaZulu-Natal. These new parties joined F. W. De Klerk’s National Party (NP), the liberal Democratic Party (DP) and the new Freedom Front (FF)—a descendant of the ‘white right’ parties of the old constitutional dispensation—in battling for the votes of millions of newly-enfranchised people.

Elections were conducted under List PR with half the National Assembly (200 members) being chosen from nine provincial lists and the other half being elected from a single national list. In effect, the country used one nationwide constituency (with 400 members) for the conversion of votes into seats, and no formal threshold for representation was imposed.

The Droop Quota was used to allocate seats, and surplus seats were awarded by an adaptation of the Largest Remainder Method. Early drafts of the electoral law put the threshold for parliamentary representation at 5 per cent of the national vote but, in a concession to the smaller parties, the ANC and the NP agreed in early 1994 to drop any ‘mandatory’ threshold. However, only those parties with 20 or more MPs, 5 per cent of the Assembly, were guaranteed portfolios in the first government’s cabinet of national unity.

The fact that the ‘Mandela liberation-movement juggernaut’ would have won the National Assembly elections under almost any electoral system cannot diminish the importance of South Africa’s choice of a List PR system for these first elections. The PR system, as an integral part of other power-sharing mechanisms in the new constitution, was crucial to creating the atmosphere of inclusiveness and reconciliation which precipitated the decline of the worst political violence and has made post-apartheid South Africa something of a beacon of hope and stability to the rest of troubled Africa.

Nevertheless, in 1990, upon Nelson Mandela’s release from prison, there was no particular reason to believe that South Africa would adopt PR. The ‘whites-only’ Parliament had always been elected by an FPTP system, while the ANC, now in a powerful bargaining position, expected to be clearly advantaged if FPTP were maintained. As only five electoral districts, out of over 700, had white majorities, the ANC, with 50–60 per cent of the popular vote, expected to win 70 per cent or 80 per cent of the parliamentary seats easily due to the vagaries of FPTP voting. But the ANC did not opt for this course because it realized that the disparities of a ‘winner-takes-all’ electoral system would be fundamentally destabilizing in the long run for minority and majority interests. List PR also avoided the politically charged and controversial question of having to draw constituency boundaries and, furthermore, it fitted in with the executive power-sharing ethos which both the ANC and the Nationalists saw as a key tenet of the interim constitution.

It is probable that, even with their geographical pockets of electoral support, the Freedom Front (which won nine seats in the new National Assembly), the Democratic Party (seven seats), the Pan-Africanist Congress (five seats), and the African Christian Democratic Party (two seats) would have failed to win a single parliamentary seat if the elections had been held under a single-member district FPTP electoral system. While these parties together only had 6 per cent of the members of the new Assembly, their importance inside the structures of government far outweighs their numerical strength.

A reading of the detailed results reveals, somewhat surprisingly, that in 1994 List PR may not have particularly advantaged the medium-sized NP and the IFP over and above the number of seats they would have expected to win under an FPTP system. This was primarily due to the ‘national referendum’ nature of the campaign, which led to a two-party battle between the old and the new—the ANC versus the IFP in KwaZulu-Natal province, and the ANC versus the NP in the rest of the country. Furthermore, the ethnically homogeneous nature of constituencies and the strong geographical concentrations of support in South Africa meant that the NP and the IFP would have won only slightly fewer seats under a constituency system. However, FPTP would in all likelihood have given the ANC a small ‘seat bonus’, increasing its share of the seats in the National Assembly beyond its share of the popular vote (which was 62 per cent) and beyond the two-thirds majority needed to draft the new constitution without reference to other parties.

The practice of having one ballot for the National Assembly and one for the provincial parliament also proved to be an important innovation in the electoral system design. Until a few months before the election, the ANC was still insisting on a single ballot which would be counted for both the national and provincial elections. This was quite clearly a manoeuvre to advantage the larger, nationally-based parties and was only changed through the pressure of an alliance of business leaders, the Democratic Party, and international advisers. The eventual results did show that large numbers of voters had split their national and provincial ballots between two parties, and it appears as though the major beneficiaries of separating the ballots were the small Democratic Party and the Freedom Front. Both polled more than 200,000 votes in the provincial elections over and above their national result, which went a long way to explain the 490,000 drop between the NP’s national and provincial totals.

The choice of electoral system has also had an impact upon the composition of the Parliament along the lines of ethnicity and gender. The South African National Assembly sworn into office in May 1994 contained over 80 former members of the whites-only parliament, but that was where the similarities between the old and the new ended. In direct contrast to South Africa’s troubled history, black sat with white, communist with conservative, Zulu with Xhosa, and Muslim with Christian. To a significant extent the diversity of the new National Assembly was a product of the use of List PR. The national, and unalterable, candidate lists allowed parties to present ethnically heterogeneous groups of candidates which, it was hoped, would have cross-cutting appeal. The resulting National Assembly was 52 per cent black (including Xhosa-, Zulu-, Sotho-, Venda-, Tswana-, Pedi-, Swazi-, Shangaan- and Ndebele-speaking), 32 per cent white (English- and Afrikaans-speaking), 8 per cent Indian, and 7 per cent Coloured—this compared to an electorate which was estimated to be 73 per cent black, 15 per cent white, 9 per cent Coloured, and 3 per cent Indian. Women made up 27 per cent of MPs.

In 1999 the proportion of black MPs rose to 58 per cent and that of Coloured MPs rose to 10 per cent, while whites made up 26 per cent and Indians 5 per cent. In 2004 the black proportion (65 per cent) came closer to their population share, while whites made up 22 per cent. Numbers of Coloured and Indian MPs held roughly steady. The proportion of women MPs rose to 30 per cent in 1999 and to 33 per cent in 2004. There is a widespread belief in South Africa that if FPTP had been introduced there would have been far fewer women, Indians and whites, with more black and male MPs.

Finally, more polarized forms of representation would be expected under FPTP, with whites (of different parties) representing majority white constituencies, Xhosas representing Xhosas, Zulus representing Zulus, and so on. While problems with lack of district accountability and of remoteness are perceived effects of the present South African List PR system, it has meant that citizens have a variety of MPs to approach when the need arises. Nevertheless, there is a continuing debate in South Africa about how to increase democratic accountability and the representativeness of the MPs. It was widely accepted that the first non-racial election was more of a referendum about which parties should draw up the new constitution. But subsequent elections have been about constituting a representative Parliament, and many political actors and voters argue that the electoral system needs to be altered to take this into account.

Today, all the major political parties still support the principle of PR. Without greatly increasing the difficulty of the ballot, voters could be allowed to choose between candidates as well as parties, without the PR character of the Parliament being affected in any way. One option is to elect MPs in smaller multi-member constituencies in order to develop a stronger geographical tie between electors and their representatives. At the moment the regional lists represent areas so large that any form of local advocacy is entirely lost. A second option is to adopt the MMP system, with half the members elected from single-member districts while the other half come from compensatory PR lists. Both these options were considered by a 12-member Task Team, led by Frederick van Zyl Slabbert, a former leader of the Democratic Party, and briefed to consider reform options in 2002. This Task Team had an inbuilt ANC- Independent Electoral Commission (IEC) majority, and was appointed by the president to review the electoral system in the light of complaints that the List PR system did not include adequate geographical representation. It ultimately recommended that South Africa should retain its List PR system but change it to a two-tier system, splitting the country into 69 constituencies electing between three and seven MPs, and keeping 100 seats as ‘compensatory’ national seats. However, the ANC government rejected this reform for the 2004 general election.

South African closed List PR ballot paper.

ACE Electoral Knowledge Network

Personal tools

Electoral Systems Case Studies

Argentina