|Staff of the Parliamentary Research Branch
(PRB) of the Library of Parliament work exclusively for Parliament conducting research and
providing analysis and policy advice to Members of the Senate and House of Commons and to
parliamentary committees on a non-partisan and confidential basis. The documents on
this site were originally prepared for general distribution to Canadian Parliamentarians
to provide background and analysis of issues that may arise in the course of their
Parliamentary duties. They are made available here as a service to the public. These
studies are not official Parliamentary or Canadian government documents. No legal or other
professional advice is offered by the authors or the Parliamentary Research Branch in
presenting its publications or in maintaining links to other Internet sites.
TABLE OF CONTENTS
B. Proportional Representation
The subject of electoral systems, though dry-as-dust to some and arcane to many, is --- or ought to be --- of central interest to anyone concerned with the operation of democratic systems of government. Elections are the defining moment in any democracy and in representative democracies, as one scholar reminds us, elections perform two fundamental tasks: they confer authorization upon those chosen to represent the electors and they hold representatives to account for their actions while in office.(2) Electoral systems, the means by which elections are formally structured, are thus a vital component in the achievement of these goals.
Strictly defined, electoral systems are the mechanisms by which the preferences of citizens are translated into seats in representative institutions. As such, their impact on a whole range of elements that make up the political character of a society is quite considerable. The behaviour of political parties and candidates for elected office will, for example, in large measure be conditioned by the shape of an electoral system. Canada's electoral system is a case in point. Academic observers have noted that in this country the electoral system is weighted in favour of regional preferences, so that parties are encouraged emphasize regional rather than national concerns during election campaigns. This has tended to produce patterns of representation at the national level whose focus is regional and whose consequences are centrifugal.(3)
More importantly, the way in which an electoral system translates votes into seats in elected assemblies may influence the degree of public support for the democratic system itself. If, for example, citizens do not perceive that their preferences are adequately reflected in the legislature following an election, their support for the system in general is likely to decline. Turnout during elections will drop off, respect for politicians and elected representatives will fall, and laws enacted by government will not be seen as fully legitimate.
Democracy is an ideal, an abstraction which often assumes concrete dimensions for the vast majority of people through the electoral system. For many, an election marks the only occasion of any form of political participation --- it is the only tangible evidence of what it means to be a citizen in a democratic society. It is therefore of utmost importance that electoral systems be seen as fair and as fulfilling public expectations; if not, democracy itself is put at risk.
Canadians are no strangers to debates over electoral systems.(4) The most recent of these debates took place during the last round of constitutional talks and focused on proposals to transform the Canadian Senate into an elected body.(5) Although there is little evidence that Canadians are ready to adopt new means of structuring their electoral systems, it is likely that these debates will continue. As the quotation that opens this paper suggests, citizens of our country are not altogether happy with the status quo. Citizen discontent with "politics as usual" has sometimes prompted a demand for changes in the structure of political institutions, including the way elections are organized, as this paper will show. The citizens of two countries, New Zealand and Italy, have just voted in large numbers for major overhauls to the structure of their electoral systems.
In view of the fact that electoral systems will remain a world-wide topic of discussion, this paper aims to acquaint the reader with the general outlines of the principal electoral systems currently used in representative democracies. The first section describes the mechanics of these systems, while the second section summarizes their claimed advantages and drawbacks. Two concluding sections examine the recent efforts by citizens of New Zealand and Italy to change their electoral systems.
The electoral systems currently in use in representative democracies can be divided into two basic kinds: majoritarian systems and proportional representation systems (often referred to as PR).
In majoritarian electoral systems, winning candidates are those having attracted the most votes in a given electoral district. Majoritarian systems differ according to the number of representatives elected in an electoral district and the kinds of majorities (simple or absolute) that winners must achieve.
Single-member plurality (SMP) systems are commonly found in countries that have inherited elements of the British parliamentary system; it is this kind of electoral system that is most familiar to Canadians.
In electoral districts represented by one member in an elected assembly, simple rather than absolute majorities suffice to determine the winner of an electoral contest.(6) Each elector marks a single "X" (or other similar mark) beside the name of the candidate of his or her choice. Although several candidates may compete for the seat, the winner need only attract the largest number of votes cast. For this reason, this kind of electoral system is referred to as a "single-member plurality" or a "first past the post" system. Electoral systems of this sort are used in Canada, the United States, New Zealand and the United Kingdom.
In some majoritarian electoral systems, more than one member per electoral district can be sent to the assembly.(7) Voters in this kind of system mark off as many names on their ballots as there are seats to be filled. As is the case in single-member systems, the candidates with the most votes are declared elected.
In contrast to the systems just described, single-member majoritarian systems seek to ensure that the winning candidate has the support of an absolute majority of the voters in his or her district. There are essentially two ways of achieving this outcome.
Countries that employ an alternative vote (also referred to as a preferential voting system or PV) require voters to rank-order their preferences on their ballots. Electors write number 1 down beside their first choice, 2 beside the second, and so on. If, when the ballots are tallied, no candidate receives an absolute majority, the candidate with the least votes is eliminated and his or her ballots are redistributed according to the second choices marked on them. This process continues until a winner emerges with more than half of the total vote. The alternative vote system has been used for elections to the Australian House of Representatives (Australia's lower house) since 1918.
The two-ballot or second-ballot system is another means of ensuring that the winning candidate is supported by a full majority of voters. Under this system, balloting may take place in two stages. During the first, voters have a choice among several candidates, only one of whom they may vote for. If no clear winner emerges from this first round of voting, a second ballot is held between the two candidates with the best showing. (In a variant of this system, when more than two candidates appear on a second ballot, a simple plurality determines the winner.)
This system was most recently used in France for parliamentary elections in March 1993. The French changed this system briefly, in 1986, replacing it with proportional representation (see below), but restored the two-ballot system shortly afterward. The system is widely used for presidential elections, including those held in France.
The second major category of electoral system is known as proportional representation or PR. PR systems are specifically designed to allocate seats in proportion to votes, in the hope that assemblies and governments will accurately reflect the preferences of the electorate. PR systems are now the most frequently used electoral systems in western democracies.(8)
Under PR, political parties are assigned a number of seats in parliament corresponding to the degree of support they have received in a given electoral district; of necessity, this arrangement dictates that all PR systems rely on multi-member districts.(9) PR systems are of two basic types: party list systems and single transferable vote (STV) systems.
Under party list systems, voters in an electoral district choose from among slates of candidates put forward by the various parties contesting an election. When the votes are tallied, each party is entitled to seat the number of members from its list that corresponds to its share of the popular vote; for example, if a given party obtains 30% of the vote, then it would send 3 members out of a list of 10 candidates to the legislature.
To discourage the emergence of splinter parties, jurisdictions using this system sometimes set a vote threshold that parties must obtain in order to qualify for seats. Thresholds vary from jurisdiction from jurisdiction; in Israel parties must receive a minimum of 1% of the popular vote in order to qualify for seats in the Knesset.(10) In Germany, on the other hand, parties must win a minimum of 5% of the national vote or win seats in three single-member constituencies before they are given seats on a proportional basis.(11) Once the threshold is met, different methods, described below, are used to calculate how seats will be allocated among the parties.
After a party's share of the available seats has been determined, it must be decided which candidates on its list will be declared elected. In most jurisdictions this depends on the order in which candidates' names appear. Those whose names appear at the top of their party's list of candidates thus have the best chance of being elected, those at the bottom, the least.
List systems have been criticized because they place considerable power in the hands of political parties, since they determine the order in which candidates names appear on the ballot --- and thus which candidates are most likely to assume office. As a consequence, voters are deprived of a significant degree of choice and the ballot itself is rendered less meaningful. In order to redress this imbalance, some jurisdictions that use a list system allow electors a greater degree of choice among the candidates. This variation on the party list system can take two forms. Under the first, voters have a choice among candidates, but they must be from one party alone. Under the second form, known as panachage (used in Switzerland), voters are permitted to make their choices regardless of party. The basic principles of the party list system are still operative, however: parties are allocated seats on the basis of the popular vote they receive.
In all PR systems, there has to be some means of determining the allocation of seats among those contesting the election. Three formulae are commonly used to do this: the largest remainder system and two highest average systems (the d'Hondt version and the Sainte-Laguë version). Table 4 provides a overview of these methods.
Under this system, the first step is to set the quota or threshold of votes that each party must attain to win a seat. The vote for each party is then divided by the electoral quota. The simplest method of establishing a quota (the "Hare quota") functions as follows: the number of votes cast is divided by the number of seats to be filled. For example, in a constituency where five seats are to be filled, and 40,000 votes are cast, the quota would be 8,000 votes. There are three other options that can be used. Under the "Hagenbach-Bischoff quota," the number of votes cast is divided by the number of seats to be filled plus one. When the "Droop quota" is used, the number of votes cast is divided by the number of seats plus one and one is added to the quotient. The "Imperiali quota," used in Italy, divides the number of votes cast by the number of seats plus two.
Subsequently, a seat is awarded to each party for each bloc of votes equal to the quota. For example, if a Hare quota has been set at 5,000 votes, each party will be given one seat for each bloc of 5,000 votes that it has received. In a second step, any remaining seats to be allocated are awarded on the basis of which party or parties have the highest number of votes remaining after the quotas have been used up. Table 1 illustrates how this works.
Four-member constituency, 20,000 votes
Source: Dick Leonard and Richard Natkiel, World Atlas of Elections: Voting Patterns in 39 Democracies, The Economist Publications, London, 1986, p. 2.
As the table shows, only parties A and B achieve the electoral quota and, as a result, only two of the four seats can be directly allocated. In the second step, the third seat goes to party A, since it has the largest number of votes remaining after the quota has been used. The party with the second-largest number of remainder votes, party C, is awarded the fourth and final seat. The largest remainder system is known to reward smaller parties; here, party C wins as many seats as party B, though it received only half as many votes.
The highest average system divides each party's votes by successive divisors and then allocates seats to the parties in descending order of the quotients. Table 2 shows the same results as Table 1 but using the d'Hondt highest average system to allocate the seats.
Four-member constituency, 20,000 votes
Source: Dick Leonard and Richard Natkiel, World Atlas of Elections: Voting Patterns in 39 Democracies, The Economist Publications, London, 1986, p. 3.
In this example, the number of votes received by each party is successively divided by d'Hondt divisors (1,2,3). Seats are allocated once the use of all the divisors has been completed; in this way it is possible to compare the quotients and allocate the seats on the basis of their descending order. Party A, with the highest quotient of 8,200, is awarded the first seat; its third-highest quotient of 4,100 gives it the third seat as well. Party B's second highest quotient of 6,100 gives it the second seat and its quotient of 3,050 gives it the fourth, and last, seat. It is clear from this example that the d'Hondt system tends to award seats to parties that receive the largest share of the votes cast, a factor which indicates that this system does not provide a large measure of proportionality.
The Sainte-Laguë and Saint-Laguë modified systems use different divisors from those used by the d'Hondt system. The Sainte-Laguë system divides a party's votes by 1,3,5,7, etc., instead of by 1,2,3,4, etc. A modified version of this system is used in several Scandinavian countries and involves setting the first divisor at 1.4 instead of 1 (so that the divisors will be 1.4, 3, 5, 7, etc.). This tends to favour medium-sized parties in a multi-party system.
It is possible to combine features of largest remainder and highest average systems by allocating seats using the Hare quota and d'Hondt divisors. In the initial stage, the quota is applied; remaining seats are then allocated through the use of the divisors.
In contrast to party list systems, STV systems emphasize the individual candidate rather than the party. As is the case in all PR systems, electoral districts using STV are represented by several members in an assembly. Voters are asked to rank-order their choices among the candidates whose names appear on the ballot. When the ballots are counted, the first step is sorting them according to the first choices. In order to be declared elected, a candidate must obtain a certain threshold of the votes cast. Those candidates who obtain the threshold during the first stage of counting are declared elected and any votes they have received in excess of the threshold are redistributed according to the second choices as marked. The second stage of counting involves the redistribution of these "surplus" ballots; once more, those candidates achieving the threshold are declared elected and any surplus votes redistributed. This process continues until all vacant seats have been filled.
A hypothetical case shows how the STV system, using a Droop quota, works. If a 3-member electoral district had 1,000 voters, the number of votes required to win would be
Three-member constituency, 1,000 votes
* = declared elected
Source: New Zealand, Electoral Referendum Panel, The Guide to the Electoral Referendum, Wellington, 1992, p. 9.
Table 3 shows how the votes would be tallied. At the first count, candidate D obtained the threshold of 251 votes, was declared elected and had his or her 19 surplus votes redistributed. These surplus votes were redistributed according to the second choices marked on them, giving the new totals, which appear in the second count column. No candidate (apart from D) reached the threshold on the second count so candidate C, who had the least votes, had all of his/her ballots redistributed for the third count: as a consequence candidate A was declared elected. On the fourth count, A's surplus ballot's were redistributed, giving E the necessary threshold. Thus after three redistributions and four counts, the three seats were filled by candidates A, D and E. It is noteworthy that candidate B, who received the second highest number of first preference votes, was not elected in the final outcome.
The STV system is currently used for elections to the Australian Senate, in Malta and in the Republic of Ireland. In Australia, voters are now offered a choice on their ballots between choosing a party list or ranking preferences among multiple candidates.
Some jurisdictions have chosen to use a mixture of majority and proportional representation systems in order to achieve the benefits of both. Since the late 1940s in Germany, for example, one half of the seats in the Bundestag (the lower house of parliament) have been filled by plurality, using single-member constituencies, while the other half are filled using party lists, according to the d'Hondt system. Voters mark two choices on their ballot papers: one from among a list of parties, the other from among a slate of candidates for district representation.
Quotas and Divisors: the Formulae
In summary, when largest remainder systems are used, quotas establish a threshold of votes that parties must attain to become eligible for seats in multi-member electoral districts. When highest average systems are used, divisors provide a means of allocating seats among the parties. Sometimes the two systems are combined so that quotas are used in the first stage and divisors are used to determine subsequent allocations.
No electoral system is perfect; each has its advantages and disadvantages of which the careful observer should be aware. This section of the paper summarizes the arguments made by proponents and critics of the various systems.
It is helpful to remember that these arguments are sometimes polemical; an objective effort to assess each system on its merits is therefore worthwhile. We should ask whether or not given electoral systems are effective in achieving certain desired outcomes. One scholar has suggested that elections in representative democracies should ideally accomplish the following goals:
One might ask, as well, whether a give electoral system is capable of achieving these aims:
However desirable, it would be next to impossible to find an electoral system that is able to satisfy all of the points listed above. Nevertheless, these criteria do provide an objective scale which can facilitate comparison and evaluation of electoral systems. The lists can also serve as a means of sorting out competing, compelling, and sometimes confusing arguments made on behalf of the various systems.
Plurality and majoritarian systems offer several benefits. The first is their ability to produce single-party, majoritarian government. Indeed, research shows that single-party rule is more likely under plurality than under PR systems.(14) This means, in effect, that when a voter votes for a party or a candidate, he or she is also choosing a government. Coalitions are less likely under this kind of electoral system and it could be argued that the governments formed as a result have a freer hand in enacting the policies on which they campaigned.
The single parties that form governments under the SMP system tend to hold legislative majorities.(15) This, in turn, tends to produce governments that are more stable (ie., are not subject to defeat in votes of confidence and can complete their legislative terms) and more capable of enacting their legislative programs. Nevertheless, SMP systems are not an absolute guarantee of majoritarian governments. It can be argued, furthermore, that majoritarian governments are not necessarily desirable.(16)
The propensity toward single party, majoritarian governments, may have an unexpected effect. Discouraged from electoral participation, small groups holding extreme positions may be prompted to resort to other than democratic means to advance their cause. If, on the other hand, the electoral system allows such groups an opportunity for parliamentary representation, a measure of conformity with established rules is imposed upon them.
A second major advantage of majoritarian systems is their relative simplicity in the eyes of the electorate. No complicated formula is involved, just the straightforward proposition that the candidate who gets the most votes wins.
There are those, however, who refute the alleged benefits of plurality and majoritarian systems. One analyst, using the means of analysis suggested above, writes that they:
The most prevalent argument is that representation is not well served by this kind of electoral system. Because majoritarian systems tend to distort outcomes by favouring strong parties and under-representing weaker ones, the wishes of most voters are often not reflected in electoral outcomes. Critics point out that this discredits the entire political system in the eyes of those it is meant to serve. At the least, citizens become uninterested in political involvement, evidenced by declining turnout at elections; at worst, they use less passive means to show dissatisfaction, so that democracy is placed at serious risk.
Since the party with the most votes generally wins more seats than its share of the popular vote would indicate, other parties are correspondingly disadvantaged by majoritarian electoral systems. As Canadian political scientist William P. Irvine explains,
The corollary is that, by exaggerating the strength of the government party, majoritarian systems produce weak, ineffective oppositions.(19)
In large countries with dispersed populations, plurality systems tend to discriminate against parties whose support, though national, is thinly spread. Parties with a strong regional presence but little national support, tend, on the other hand to be rewarded. A majoritarian electoral system may thus be viewed as a divisive factor in a country with distinct regions, since parties have incentives (in terms of electoral rewards) to make strong regional appeals. Those whose support is concentrated in one region, or who become champions of regional rather than national concerns, will be rewarded and regional cleavages will be consequently enhanced.
Variants of majoritarian and plurality systems have attracted the following comments:
Single-member-plurality systems are considered by many to be superior in their representation of constituency interests. In majoritarian and plurality systems that elect one representative per district there is a direct connection between elected representatives and electors. There can be no ambiguity over who is responsible for a constituency's interests, unlike the case with PR systems where districts send more than one representative to the legislature.
These systems combine some of the advantages of PR and plurality systems by ensuring that not only do multiple points of view within a constituency gain legislative representation (as is the case under PR), but also that those elected have significant levels of support from the electorate. Because each district is represented by several members, however, the lines of responsibility between the elector and the elected are not always clear.
Single-member majoritarian systems, in contrast to SMP systems, are designed to ensure that elected representatives have the support of a full majority of the constituency's electors. A government elected under this system thereby enjoys enhanced legitimacy.(20) There is, however, no guarantee that the party or parties who form the government under this system have the support of a clear majority of all voters.(21)
An additional argument in favour of this kind of electoral system is its effect on extremist movements. Advocates claim that by favouring strong parties majoritarian systems discourage the emergence of extremist movements. Evidence tends to confirm this claim.(22)
Lastly, under the two-ballot form of single-member majoritarian systems, voters need be less concerned about "wasting" their votes on the first ballot, since they know they will probably be given a second choice.
The principal argument advanced in favour of PR is its ability to reflect more accurately the preferences of voters in terms of seats in parliament. Voters are said to be more willing to cast votes for smaller parties when they know that their votes will produce tangible results, and when seats are allocated on the basis of the share of the popular vote. The ability, in general, for PR systems to deliver seats to smaller parties encourages the formation of such parties, a factor which promises representation of a wider spectrum of public opinion.
While studies confirm that PR systems on average offer greater proportionality than majoritarian/plurality systems,(23) some researchers have discovered that it is still possible for a plurality system to produce a more proportional result than a PR system.(24) Other factors, such as the number of parties contesting an election, may influence the proportionality of electoral outcomes.(25)
A closely related argument is that PR systems offer greater opportunities for legislative representation for minority groups and women. However, a recent Canadian Royal Commission on electoral reform pointed out that levels of women's representation in elected assemblies are often attributable to variables other than the electoral system, for example political parties' adoption of quotas for women candidates. In this respect, the behaviour of political parties, especially in party-list PR systems, is crucial to women's ability to gain seats in elected assemblies. More careful analysis reveals that PR jurisdictions that do not use a quota system for women candidates have records similar to Canada's in this regard -- and sometimes worse.(26)
It is also claimed that, because minority views are not marginalized, political discourse and political participation are enlivened in PR systems. The one reliable empirical indicator for this assertion, levels of voter turnout at elections, tends to confirm this.(27)
Arguments against PR, however, can be just as compelling as those in its favour. Some critics point out that PR systems encourage the emergence of extreme views, which, though quite often based on short-lived opinions of the day, are given a certain longevity and enhanced legitimacy through access to parliamentary representation. This argument is best summed up by Irvine, who writes that under PR systems
PR systems are also criticized for the complexity of their balloting process and the way in which votes are tallied. Available information suggests that, while voter turnout may indeed be high in PR systems, ballot spoilage is also high, a possible sign of voter confusion when offered a multiplicity of choice.(29)
Concerns are also expressed about the shape of the governments that result from PR electoral systems. In contrast to plurality/majoritarian systems which favour the formation one-party governments, PR systems generally produce coalitions. Thus, when casting their ballots in a PR system, voters are not electing a government. Governments under PR are typically formed after elections, when parties attempt, through a bargaining process, to build governing coalitions. Voters, in effect, have little direct say regarding the complexion of their government.
As well, coalition governments are viewed as less than stable. The bargaining among parties continues after the government-building process as various elements strive to have parts of their agenda adopted as official policy. Compromise can be brokered, but negotiations often produce rifts that cannot be resolved. Consequently, unless there are changes in governing coalitions, the coalition will collapse, leading to new elections. It is on this basis that some have argued that coalition government is inherently unstable so that the electoral system that produces it is unsatisfactory.
Critics also claim that, contrary to appearances, coalitions actually make it more difficult to change governments. Coalition membership may fluctuate following elections, but the stronger members usually remain in place.(30) As a consequence, it is much more difficult to change a government under PR than under plurality/majoritarian systems.
Additional claims are made with respect to the advantages and disadvantages offered by specific forms of PR.
Those who defend party list PR systems argue that political parties occupy an important place in any representative democracy and that the list system helps to ensure that the role of parties will be maintained and strengthened.
On the other hand, parties acquire too much power when they can determine whose names will appear at the top of the lists. Those elected on the basis of this system owe primary allegiance to their parties rather than to their electorates. Such concerns can be addressed by allowing voters to choose among lists or within lists.
Variants of the party list system that allow voters to chose either from among a party's candidates or between the candidates on several parties' lists allow voters a truer expression of their preferences. These variants help ensure that the candidates with the strongest levels of support are elected; the legitimacy of the outcome is enhanced.
If however, voter choice is restricted to the candidates from a single party, candidates will be encouraged to compete against members of their own party rather than those from rival formations. This puts party cohesion at risk and makes the task of governing more difficult.
Because STV, unlike party-list forms of PR, emphasizes the candidate rather than the party, it offers an advantage in that it "...makes possible the fair representation of opinions which do not coincide with party divisions."(31) Also, in contrast with party list forms of PR, representatives under STV systems are more inclined to be attentive to constituency needs because they depend on their constituents, rather than their party, for re-election.
Critics of STV claim that it leads to weaker parties and hinders the emergence of a responsible party system because candidates work to attract personal support, sometimes at the expense of other candidates from their own party.(32)
Experience in Australia has shown that when as many as fifty candidates can contest an election under an STV system, the process of counting the vote may be lengthy. An additional disadvantage, also drawn from Australia, is the complexity of casting a ballot, as suggested by high levels of ballot-spoilage.(33)
Two countries, New Zealand and Italy, have recently held referendums in which citizens expressed a desire to adopt new electoral systems. New Zealanders indicated a willingness to move away from SMP towards an arrangement including some proportionality, while Italians went in the opposite direction. It is noteworthy that in both instances the decision to change systems took place against a background of dissatisfaction with "politics as usual."
On 18 September 1992, New Zealanders voted in a referendum on electoral change. Voters were asked whether they supported their single-member plurality system and, if not, which of four different PR models they would prefer. Approximately 85% of those who cast ballots voted for a change of system and of these an overwhelming 70% favoured the mixed system used in Germany (see above). September's referendum results are not binding; New Zealanders will have an opportunity to choose between the status quo and a refined version of the mixed system during the next general election, expected sometime in late 1993.
In a recently held referendum (18 April 1993), Italians voted to leave the PR electoral system, which has given them 51 governments since the end of the Second World War. Voters were asked to respond to questions dealing with party financing and the system of proportional representation used for elections to the upper and lower houses of parliament. The decision will almost certainly produce a change in the electoral system; approximately two-thirds of Italy's 48 million eligible voters cast ballots, many more than the turnout required to make the outcome binding (50%).(34)
Italians voted to move to a single-member plurality system, which, it is hoped, will produce single-party governments and put an end to the power of party bosses. Through their votes, Italians have approved a plan to fill three quarters of their Senate's seats (238 out of the 315 seats) by the first-past-the-post system. It is hoped that the same rules will be subsequently applied to the Italian lower house. It will be the government's task to determine the eventual shape of electoral reform.
While electoral systems are a vital component of any representative democracy, one should not overstate their importance. Even the best electoral system will fail if other conditions are not met. Two Canadian political scientists have written a thoughtful reminder of this. "Electoral systems," they write,
As stated in the introduction to this paper, elections constitute the most direct, and indeed for many the only, experience of what it is to be a citizen in a democratic society. Perhaps it is for this reason that rising levels of discontent result in demands for change in the way elections are structured, as was seen recently in New Zealand and Italy, when a change of electoral system appeared to be a convenient and effective way of redressing the wrongs in the democracies as a whole. Canadians, beset with many of the same doubts and discontents, may well begin to ask whether they desire a similar change.
Blais, André. "The Debate over Electoral Systems." International Political Science Review, Vol. 12, No. 3, 1991, p. 239-260.
Blais, André and Elizabeth Gidengil. Making Representative Democracy Work: The Views of Canadians. Vol. 17 of the Research Studies. Royal Commission on Electoral Reform and Party Financing, Toronto, 1991.
Freedom House Survey Team. Freedom in the World: Political Rights and Civil Liberties 1991- 1992. Freedom House, New York, 1992.
Interparliamentary Union. Parliaments of the World. Vol. I. 2nd edition. Facts on File Publications, New York, 1986.
Irvine, William P. Does Canada Need a New Electoral System? Institute of Intergovernmental Affairs, Queen's University, Kingston, Ontario, 1979.
Jackson, Robert and Doreen. Politics in Canada. 2nd edition. Prentice-Hall Canada, Scarborough, Ontario, 1990.
Lakeman, Enid. How Democracies Vote: A Study of Electoral Systems. Faber and Faber, London, 1974.
Leonard, Dick and Richard Natkiel. World Atlas of Elections: Voting Patterns in 39 Democracies. The Economist Publications, London, 1986
Mackie, Thomas T. and Richard Rose. The International Almanac of Electoral History. 3rd edition. Macmillan, London, 1991.
New Zealand, Electoral Referendum Panel. The Guide to the Electoral Referendum. Wellington, 1992.
Rose, Richard, editor. Electoral Participation: A Comparative Analysis. Sage Publications, Beverly Hills, 1980.
(1) André Blais and Elizabeth Gidengil, Making Representative Democracy Work: The Views of Canadians, Vol. 17 of the Research Studies, Royal Commission on Electoral Reform and Party Financing, Toronto, 1991, p. 79. (2) Hanna Fenichel Pitkin, The Concept of Representation, University of California Press, 1967, esp. p. 38 - 59. (3) See, for example, Alan Cairns, "The Electoral System and the Party System in Canada, 1921-1965," Canadian Journal of Political Science, Vol. 1, No. 1, March 1968, p. 55-80 and William P. Irvine, Does Canada Need a New Electoral System? Kingston, Ontario, 1979. (4) See, for example, William P. Irvine, "A Review and Evaluation of Electoral Reform Proposals," in Peter Aucoin, editor, Institutional Reforms for Representative Government, Volume 38 of the research studies, Royal Commission on the Economic Union and Development Prospects for Canada, University of Toronto Press, Toronto, 1985, p. 71-109. Irvine provides an overview of the various kinds of electoral systems proposed for use in Canada. (5) The Special Joint Committee on a Renewed Canada (Beaudoin-Dobbie Committee) recommended that the Senate be elected on the basis of proportional representation. The later Charlottetown Accord agreed only that the Senate would be elected (either by the populations of provinces or by their elected assemblies) without specifying any electoral system. (6) Simple majorities are pluralities or most of the ballots cast. Absolute majorities are defined as 50% plus 1. (7) In Canada, for example, Prince Edward Island has a 32-member provincial assembly elected from 16 ridings. (8) Thomas T. Mackie and Richard Rose, The International Almanac of Electoral History, 3rd edition, Macmillan, London, 1991, p. 503. (9) Note that in some instances an entire country can be considered as a single electoral district for the purposes of an election held under PR. This is the case in Israel. (10) Israel's low threshold has been criticized because it fosters a multitude of small parties, most of which have no difficulty in winning seats. The result is coalition government in which small fringe parties may exercise an influence out of proportion to the support they have received during elections. (11) The arrangements in Germany were put in place as a direct consequence of Germany's experience leading up to the Second World War; the high thresholds were designed to discourage the emergence of small extremist parties. (12) Enid Lakeman, How Democracies Vote: A Study of Electoral Systems, Faber and Faber, London, 1974, p. 28 . (13) New Zealand, Electoral Referendum Panel, The Guide to the Electoral Referendum, Wellington, 1992, p. 4. (14) André Blais, "The Debate over Electoral Systems," International Political Science Review, Vol. 12, No. 3, 1991, p. 240. (15) Ibid., p. 241. On the basis of earlier research, Blais reports that 72% of single-member district plurality elections produce a one-party legislative majority, compared to 10% of PR elections. He also indicates that the probability of a one-party majority government is 40 percentage points higher in a plurality than in a PR election. (16) See, for example, Cairns (1968). Cairns calls the ability of Canada's single-member plurality system to produce stable majoritarian governments "mediocre." It is argued that democracy is better served by minority governments which must take a wider spectrum of opinion into account when enacting legislative programs. (17) Lakeman (1974), p. 57. (18) William P. Irvine, Does Canada Need a New Electoral System? Institute of Intergovernmental Affairs, Queen's University, Kingston, Ontario, 1979, p. 11. (19) Cairns (1968), p. 57. (20) Blais (1991), p. 246. (21) bid., p. 247; Blais points out that the Gaullists formed the government of France in 1968 with 60% of the seats in the National Assembly, but with only 36% of the national popular vote. This result was repeated in French parliamentary elections of March 1993, when parties on the right attracted 39% of the vote and won 80% of the seats. The National Front, which was supported by 12.5% of voters, won no seats at all. (22) Ibid., Blais finds that extremist parties' share of seats is lowest in majority systems, being typically 8 points lower than in a PR system and 2 points lower than in plurality systems. (23) See, for example, Douglas W. Raie, The Political Consequences of Electoral Laws, Yale University Press, 1967, p. 96-97 and Arend Lipjhart, "Degrees of Proportionality of Proportional Representation Systems," in Bernard Grofman and Arend Lipjhart (editors) Electoral Laws and their Political Consequences, Agathon Press, 1986, p. 170. (24) Mackie and Rose (1991), devised an Index of Proportionality to assess the electoral systems in 25 countries. This index was calculated by summing the difference between each party's percentage share of seats and its share of votes, dividing by two and subtracting the result from 100. Mackie and Rose discovered that among PR systems, the index of proportionality ranged from a low of 87% (Spain) to a high of 100% (Malta). For plurality systems, the scores ranged from 79% (the United Kingdom) to 94% (the U.S. Congress). Their results showed that, on average, PR systems achieve higher proportionality than plurality systems (94% vs 86%). Individual plurality systems, however, are empirically capable (as the authors point out) of achieving higher proportionality than some PR systems. The U.S. Congress, for example, achieves a higher score than the Spanish electoral system. This may come about as a consequence of the number of parties contesting the elections, rather than the structure of the electoral system. Where plurality elections take place in a two-party system, as in the case of U.S. congressional elections, seat allocation is likely to be roughly proportional to the vote. However, when more than two parties participate in elections under a plurality system, the degree of proportionality is bound to plummet. (25) Arend Lijphart, "The Political Consequences of Electoral Laws, 1945-85" American Political Science Review, Vol. 84, No. 2, June 1990, p. 481-496. Lijphart point outs that the degree of proportionality depends on the particular quotas and divisors used in seat allocation. His analysis shows that among PR systems using the highest average method of seat allocation (which uses divisors), the Sainte-Laguë method "approximates proportionality very closely," while the modified version of the system does less well and the d'Hondt system least well. In the largest remainder systems, use of the Hare quota "tends to yield closely proportional results," while use of the Droop and Imperiali quotas produces less proportional outcomes. Overall, the d'Hondt method is judged to be the least proportional, while use of the Sainte-Laguë highest average and largest remainder system using a Hare quota produce the greatest proportionality. Lijphart bases his assessment on the degree to which small and large parties are treated evenhandedly by the formulas. (26) Canada, Royal Commission on Electoral Reform and Party Financing, Reforming Electoral Democracy, Vol. I, Ottawa, 1991, p. 20. (27) Political scientist G. Bingham Powell Jr. reports that turnout averages about 78% in PR systems and only about 71% in majoritarian systems. Powell adds, however, that some of the difference can be attributed to the facilitated registration and compulsory voting sometimes present in PR systems. See Voting Turnout in Thirty Democracies, Richard Rose, editor, Electoral Participation: A Comparative Analysis, Sage Publications, Beverly Hills, 1980, p. 12. (28) Ibid., p. 18. (29) Blais (1991), however, p. 249, observes that for democratic purposes, "...the more information the ballot reveals about voters' intentions, the better the procedure." In this respect, ballots cast under a PR system could be said to serve democratic purposes more effectively by conveying a broader spectrum of preferences. (30) For example, in Italy, the Christian Democrats have dominated all of the 51 governing coalitions formed since the end of the Second World War. (31) Lakeman (1974), p. 129. (32) Blais (1991), p. 248-249. (33) Lakeman (1974), disagrees that STV systems are too complicated for the average voter and suggests that any relationship between levels of ballot spoilage and STV is unproven. (34) Globe and Mail (Toronto), 19 April 1993, p. A7. (35) Robert J. Jackson and Doreen Jackson, Politics in Canada, 2nd edition, Prentice-Hall Canada Ltd., Scarborough, Ontario, 1990, p. 501.