Proportional representation can potentially remedy certain flaws of the plurality system. For instance, under plurality, the number of seats a party holds in the legislature often bears little relation to its share of the popular vote. This comes as a surprise to some of plurality’s advocates. The plurality system exaggerates the support for the leading party and minimizes that of other parties, leading to election results that do not mirror the popular vote. On Prince Edward Island an obvious effect of this distortion  is the virtual elimination of opposition parties from the Legislature. Recent elections have been winner-take-all affairs, resulting in exaggerated majorities for the leading party.

The plurality system allows small shifts in the popular vote to rearrange drastically the face of the government and the legislature. The “landslides” to which the press often refer are often created by these minor shifts. The Liberals carried the 1943 PEI provincial election with 20 of 30 seats; a shift of fewer than 100 votes across the province would have given the Conservatives a majority.

Electoral Politics on Prince Edward Island

Prince Edward Island’s first Assembly met, according to legend, in a Charlottetown tavern in 1773. The 18 members, elected by the male protestants of the colony, were called “a damned queer parliament” by the Sergeant-at-Arms, who was reportedly fined five shillings for the comment. The Assembly grew to 24 members by the Election Act of 1838, and to 30, in 15 dual constituencies, in 1856. The Upper House, called the Legislative Council, became elective in 1862, with six dual-member constituencies and one with a single member.

The houses were merged by the Legislature Act of 1893. Henceforth there would be a Legislative Assembly with 15 dual-member districts, each electing an assemblyman and a councillor. The function of the Assembly, as Frank MacKinnon wrote in his seminal 1951 book The Government of Prince Edward Island, was (and is) “to enable the representatives of the people to make the laws by which the province is governed, to express ideas and opinions upon public business, and to praise and criticize the actions of the executive.”

The Assembly retained the complexion provided by the Legislature Act until the 1960s. In 1962 the Royal Commission on Electoral Reform called for a revision of the voting system. The property-based franchise for electing councillors was abolished, though members continued to be designated as councillors or assemblymen. The Assembly grew to 32 members representing 16 dual-member districts.

The system changed once again after the 1994 report of the Election Act and Electoral Boundaries Commission. The Commission recommended a new electoral map, with 30 single-member districts. The Legislature opted for an alternative map proposed in a Private Member’s Bill. There are now 27 single-member constituencies. Their representatives are elected, as they have always been, by the plurality system.

While being cautious about generalizations, it is possible to say that Island political culture has been marked by partisanship and party loyalty, and that close acquaintance between MLAs and their constituents has been a normal feature of the landscape. Political partisanship was historically strong enough that “changing one’s party politics was akin to treachery or betrayal, an act of dishonour almost like changing one’s religion.”¹ The small size of the Island helped to shape this culture. Close links between voters and their representatives are encouraged by the low ratio of residents to MLAs, presently providing a population of about 140,000 with 27 representatives (nearly one MLA for every 5,000 people). Islanders typically feel little reluctance to phone their MLAs, and are likely to get through. MLAs, for their part, cultivate durable personal links with constituents.

It has been said that between the federal Parliament, the provincial Legislature, and local governments the Island possesses “perhaps more formal government than anywhere else in the world.”² The population is small – less than one-half of one per cent of the Canadian population, barely twice what it was in the 1850s – and fairly homogeneous. Post-Confederation Island politics have not, for the most part, been driven by ethnic or linguistic rivalries, though such conflicts certainly existed beneath the surface. The dual-member electoral system survived into the 1990s partly in order to accommodate religious differences.

Island elections have always been marked by disproportionality between the distribution of legislative seats and that of the popular vote. This was not a serious problem until recently, when small differences in the popular vote between the winner and loser became wide and unpredictable swings, and a viable third party, the New Democratic Party, became competitive with the Liberals and Progressive Conservatives. In 1996, for only the second time since 1923, the winning party did not win a majority of the popular vote. The Progressive Conservatives won with only 47.7 per cent of the vote.

The emergence of a third party, and the ever larger swings in the popular vote from election to election, suggest that Island political culture is changing. This makes the flaws in the plurality electoral system more visible, and more troublesome for the functioning of democracy, than ever before.

Varieties of Electoral Systems

It is a mistake to consider an electoral system a technical mechanism without influence upon day-to-day political life. The electoral system influences the outcome of every election, often decisively.

As well as affecting the way votes are translated into seats, the electoral system can influence how people vote. The Canadian plurality system, for instance, may encourage electors to vote “strategically,” for the candidate who seems to have the best chance of winning, in order to ensure the defeat of another candidate whom they oppose. When voters do this, they often do not vote for the candidate they actually support, if that person seems unlikely to win. Proportional representation, on the other hand, may encourage people to vote for small parties that are more likely to gain representation under such a system than under a first-past-the-post plurality system. It is important to be mindful of the influence wielded by an electoral system when deciding which one is most appropriate to the circumstances of a particular jurisdiction.

The designers of an electoral system must account for several important considerations. Among the most significant of these is the method of counting votes, or electoral formula. Votes may be counted by plurality (as in the first-past-the-post system), where the candidate with the most votes is elected; by majority, where the winning candidate must poll more than half the votes; or by proportionality, where several members are elected proportionally to their parties’ respective shares of the popular vote.

Another necessary consideration is the number of members elected from each constituency, the district magnitude. In plurality and majoritarian systems, as a rule, a constituency has a single member; proportional representation systems require districts with several members (the more the better, in fact, since more proportionality is possible with more members). Other considerations include the extent of choice between candidates of the same party (in proportional systems) and the form of the ballot.

Choices regarding one criterion affect other criteria; for instance, proportional electoral requires multi-member constituencies (i.e., high district magnitude) and considerable choice among candidates of the same party.

Single-Member Plurality (SMP), or First-Past-the-Post (FPTP)

This system exaggerates the strength of the strongest party, awarding it a number of seats out of proportion to its share of the popular vote. This can happen under any electoral system, but it is most pronounced under SMP. In the 1993 federal general election, the Liberals won 177 of 295 seats in the House of Commons – about 60 per cent of the total – with a popular vote of just 41 per cent. They repeated this dubious feat in 1997. Their share of the popular vote fell to 38 per cent, but the party still held a majority in the 301-seat house, with 155 seats (about 51.5 per cent).

Where the leading party polls a majority of the popular vote, not just a plurality, the magnifying effect can be grotesque. In the 1987 New Brunswick provincial election, the Liberal Party, with about 60 per cent of the popular vote, won all 58 seats in the Legislature. The Prince Edward Island Liberals won every seat in 1935 with 58 per cent of the vote.

This magnifying effect has been decisive in allocating seats in most Prince Edward Island elections of the past decade. In 1989, the Liberals won 30 seats of 32 (almost 94 per cent) with 61 per cent of the vote, and 31 seats of 32 (about 97 per cent) with 55 per cent of the vote in 1993. In April 2000 the incumbent Progressive Conservatives won just over 95 per cent of the seats (26 of 27) with about 58 per cent of the vote.

Supporters of the first-past-the-post system argue that it provides stable government by manufacturing majority governments out of minorities of the popular vote. This reflects electoral values; specifically, the conviction that creating majority governments – artificially if necessary – is more important to a polity than representing the choices of the voters. This is a choice which every democracy needs to make, but one that is not often made consciously. More often, electoral systems become institutionalized until it seems that they are somehow natural phenomena; this dampens any discussion of reform. The Canadian and PEI electorates have never actually been asked to choose an electoral system that systematically distorts their choices.

Government stability is often assumed to be an effect of the plurality system. The evidence, however, suggests no such direct link between the electoral system and political stability. The Independent Commission on the Voting System, often called the Jenkins Commission, noted, referring to the British parliament that “in only 64 of the past 150 years has there prevailed the alleged principal benefit of the FPTP system, the production of single-party government with an undisputed command over the House of Commons.”³ The record in Canadian federal elections is somewhat better, but the system is hardly efficient at manufacturing majorities; it did so on only half the occasions between 1921 and 1965 when the winning party did not have a majority of the popular vote.

Another benefit of plurality is said to be its simplicity. In pursuing this argument, there is a danger of assuming that voters are not intelligent enough to comprehend a different system than the one to which they are accustomed, and of ignoring the fact that more complex systems are used all over the world, apparently without bewildering voters.

Plurality supposedly encourages parties to compromise and discourages extremism. But again, such moderation seems likely to be as much or more a product of political tradition and culture than of the electoral system. Further, freezing out smaller parties is not necessarily synonymous with discouraging “extremism.” But keeping small parties out of the legislature can serve the interests of large established parties, which is one reason parties in power are often reluctant to tamper with the plurality system that brought them to power. Maintaining the plurality system despite the rise of smaller parties only makes matters worse, of course, since a plurality system with more than two parties has effects even more distorting than it does when dividing seats between two parties.

Finally, proponents of plurality extol the benefits of maintaining a strong link between the representative and a constituency or district. This is the most convincing defence of the plurality system. No one is eager to dispense entirely with the geographical link between the representative and the electors. Where plurality systems are reformed to be more proportional (as in New Zealand), some form of constituency representation is usually retained. A proposal to eliminate geographic constituencies would be a non-starter on Prince Edward Island, with its tradition of close links between MLAs and their constituents, and the emphasis MLAs place on constituency service. It is important to consider, however, that “there is a deep-seated conflict between the notion of territorial representation and the representation of parties”.⁴ It is not always clear whether a representative’s first loyalty is to the party or the district.

Majoritarian Systems

Majority-based electoral systems are designed to ensure that candidates are not elected without majority support. “The essential point about the rule of majorities,” writes Douglas Rae, “is that the winning party has defeated the entire field of opposition; no combination of opponents can match its numerical strength.”⁵ It is important to avoid confusing majoritarianism, which makes no promise of proportionality, with proportional representation.

There are two chief methods of creating majorities: the alternative vote and the second ballot.

In an alternative vote (AV) system, electors rank the candidates numerically on the ballot. The ballot is valid as long as one candidate is selected with the number “1.” If no one wins a majority of the votes, the candidate with the fewest first-choice votes is dropped from the ballot and the second choices indicated on his/her ballots are distributed among the other candidates. In some systems, candidates not attaining a given percentage of the first-choice votes are dropped after the first ballot. This process continues until one candidate has a majority of the vote.

While it is usually associated with Australia (sometimes under the name “preferential voting”), the alternative vote has also been used in Canada. The British Columbia elections of 1952 and 1953 were conducted under an AV system, while rural MLAs in Alberta (between 1926 and 1959) and Manitoba (1927 to 1936) were elected by AV.

In a second ballot system, if no candidate attains a majority, the weaker candidates withdraw (either by requirement, for those candidates not attracting a given minimum of the vote, or voluntarily, depending on how the system is designed). The remaining (usually two) candidates compete on a second ballot. The candidate who leads the second vote (by majority or plurality) is elected. An alternative run-off system is used in French presidential elections, where only the two leading candidates from the first round of voting are entered in the second round, to ensure that no president is elected without majority popular support.

The second ballot system has been criticized for being very disproportional – even more so than the plurality system – and expensive and time-consuming to operate.

Proportional Representation Systems

Electoral systems based on proportional representation (PR) differ from the systems considered above in that they focus first on “the principle of representation, seeking to effect the return of assemblies which, in party, social, gender, and ethnic composition, closely mirror the profile and wishes of the electorate.” By returning legislators in rough proportion to parties’ share of the popular vote, PR prevents the sort of disproportional, winner-take-all election results that the plurality system allows.

Proportional representation makes it easier for small parties to elect members where they lack the geographically concentrated support to win a seat in a constituency, yet have support throughout the jurisdiction. Thus PR is often advocated by third parties trying to break into two-party systems.⁶ The first PR system went into effect in Belgium in 1899 “because the development of a three-party system made the future relationship between votes and seats unpredictable.”

A side effect of accurately representing the electorate’s wishes is that PR produces coalition and minority governments more often than does the plurality system. Some commentators suggest that “representation versus effectiveness” is a necessary dichotomy, that stable governments and accurate representation cannot coexist. According to Maurice Duverger, by “dispersing the voters among numerous independent parties, PR prevents the citizens from expressing a clear choice for a governmental team.” Critics of PR commonly cite the example of Italy, charging that in that country PR helped create a parliament with more than 40 parties and frequent changes of government until the electoral reforms of 1993. Those critics often neglect to mention that in 1993 Italy replaced its pure party list PR system with mixed-member PR, rather than first-past-the-post. PR’s supporters point out that virtually all of the stable governments of northern and western Europe are elected by some form of PR. As Henry Milner writes, “recent experience in European countries shows that allocating seats to parties based on their popular vote need not lead to instability.”⁷ The key is to guard against an onslaught of “small, one-issue or narrowly-based parties.” This can be accomplished by setting a “threshold” – a minimum percentage of the popular vote necessary for a party to be proportionally represented in the legislature.

It has been demonstrated that PR, particularly in its party list form, tends to bring a much higher percentage of women and minority members into legislatures than SMP or majoritarian systems do. Indeed, one of the functions of Prince Edward Island’s dual-member system was to allow parties to put forward candidates of different social or professional backgrounds (e.g., religious) in the same district.

Canadian experience with PR is limited, but there is a thriving lobby among some academics, journalists and political parties for changes to the present electoral system. PR was used on a limited basis in some of the western provinces earlier in this century, and Quebec made an abortive move toward PR in the 1970s and 1980s. Canadians tend to agitate for PR after particularly appalling distortions, such as the 1980 federal election that left a liberal majority government without a seat in the western provinces, despite a healthy share of the popular vote there. The discontented political culture of the 1990s spurred another surge of interest in electoral reform. In May 2000, Members of Parliament began the first debate on PR since 1923, on a motion sponsored by New Democrat Lorne Nystrom calling for the introduction of an element of PR into the federal electoral system. In July 2000, the Institute for Research on Public Policy reported that 49 per cent of Canadians find the current electoral system unacceptable, although it did not suggest a specific alternative.

Types of Proportional Representation

The single transferable vote is the PR system most often used or advocated in Britain and its former colonies, including Ireland, Australia and Malta. The STV voting procedure resembles the alternative vote. Like the AV, the STV requires the voter to rank the candidates in numerical order. But, unlike AV districts, STV constituencies have several members, and each party usually puts forward as many candidates as there are seats to be filled. An “electoral quota” of votes needed for election is calculated, roughly by dividing the number of votes cast by the number of seats available. Any candidate whose first-place votes equal or exceed the electoral quota is elected. If seats are left unfilled, the second preferences of those ballots that elected the first member are distributed, until all the seats are filled.

The single transferable vote’s chief advantages are the high degree of proportionality it allows and the ability for voters to choose between different candidates from the same party, rather than presenting them with a single candidate or list selected by the party brass. As well, the STV wastes  fewer votes than other systems. In the Irish elections of November 1982, 83 per cent of the votes cast helped to elect a candidate; Vernon Bogdanor contrasts this figure to the nearly 70 per cent of votes that were wasted in the British constituency of Barking in 1983.

STV is the only PR system ever used in Canada. Some urban members of the Alberta Legislature were elected by STV between 1926 and 1959, as were Winnipeg members of the Manitoba Legislature between 1920 and 1953.

The Party List system has been called “potentially, the most truly representative form of proportional representation, ... being designed to return members reflecting the broadest possible spectrum of public opinion.” It demands large multi-member constituencies. In every district, each party presents a list of candidates, and seats are allocated in proportion to popular vote. Thus, if party A gains 40 per cent of the vote in a ten-seat district, it is entitled to four seats, and the first four candidates on Party A’s list are declared elected.

There may be a minimum percentage of the popular vote necessary to be entitled to seats in a list system. This threshold is designed to minimize the influence of small extremist and splinter parties. Where such thresholds are very low or do not exist, as in Israel and Italy (before it scaled back the proportionality of its system), minor parties have the potential to dominate parliaments. This is a common criticism of list PR. Opponents also say the list system breeds coalition governments, since, like any PR system, it makes majorities hard to attain. They argue further that list PR weakens the link between representatives and constituents, since constituencies are geographically large and have several representatives. Finally, they argue that list PR leaves too much power in the hands of party hierarchies, since they usually decide who goes on the list and in what order.

Mixed Electoral Systems

An electoral system need not be wholly based on PR or on plurality. Elements of proportionality and  plurality (or majority vote) can mix in a single system. As the Jenkins Commission noted, a mixed system has the benefit of flexibility; depending on the ratio and distribution between plurality and proportional seats, “varying degrees of priority can be given to proportionality on the one hand and to the constituency link on the other.” The most common form of mixed system is the mixed-member proportional (MMP) system (sometimes called the additional-member system, or AMS), as seen in Germany, New Zealand, Scotland and Wales.

Under an MMP/AMS system, each voter marks two ballots: one for a constituency member and one for a party list covering several constituencies  or the entire jurisdiction. A party’s seat entitlement accords with its proportional list vote; the number of constituency seats the party wins is subtracted from this total, and the result is the number of list seats to which the party is entitled. If the party wins more constituency seats than is eligible for based on the popular vote, it keeps the surplus and the assembly temporarily expands. As in pure party list systems, there is usually a threshold of popular vote percentage or constituency seats necessary to entitle a party to proportional representation in the legislature.

Some International Comparisons

Australia belies the myth that former British colonies inevitably develop Westminster-type electoral systems. Like Canada, Australia is a federal state; unlike Canada, Australia and its component states use a mix of majoritarian and proportional electoral systems.

Most relevant, from PEI’s perspective, is the electoral system of the state of Tasmania. Like PEI, Tasmania is a small island (with a population of about 472,000 spread across 68,000 square kilometres). Unlike PEI, Tasmania has a bicameral parliament. The upper house, the Legislative Council, is elected from 15 single-member electoral divisions by alternative vote. Tasmania uses the Single Transferable Vote for elections to the House of Assembly, the lower house, with 25 members representing five electorates. Until 1998 the House had 35 members in 7 electorates. The cut was made over the objections of the Green Party, which argued that Labour and the Liberals were trying to rig the system to diminish the parliamentary strength of smaller parties (like the Greens).

The electoral system for the German lower house (the Bundestag) combines a constituency-based plurality system (for half the seats) with a proportional party list system for the remainder. Voters cast one vote for a constituency candidate and a second vote for a regional party list (there is a separate list for each land, or state). Parties may only be represented in the Bundestag if they collect three constituency seats or five per cent of the popular vote nationally. Votes are counted in a three-step process: (i) the (list) votes are counted in order to calculate the proportional distribution of seats for the entire country. (ii) The seats are allocated between parties on a state-by-state basis. (iii) In each state, the number of constituency seats won by each party is subtracted from the party’s total seat entitlement as calculated in step (ii). The resulting figure is the number of list seats to which the party is entitled. If the party has more constituency seats than its proportional popular vote entitles it, it keeps the extra seats, which are temporarily added to the Bundestag.

The German system is usually favoured by advocates of proportionality in Commonwealth countries. New Zealand modelled its mixed system on the German example, and the electoral system for the new Scottish and Welsh parliaments follows the German pattern. Advocates of electoral reform in Canada have often proposed some form of the German model as well.

Iceland uses a party list system. The Icelandic parliament (the Althingi) has 63 members in eight constituencies, with at least five seats in each. Most of the seats in each constituency are divided among parties in proportion to their popular vote in the constituency; one quarter are distributed according to the national vote. The electoral districts are skewed against Reykjavík, the capital and largest city, which contains about 60 per cent of the population but less than half the seats. The Icelandic system has a low threshold; in some constituencies a candidate can be elected with 750 votes.

Malta, with a population of about 400,000, has a 65-member House of Representatives elected by Single Transferable Vote in 13 five-seat constituencies. Malta adopted STV in 1921, with the support of British governor George Strickland, who had been impressed with its operation in Tasmania, where he had been governor previously.

The Maltese electoral system presents several interesting features. Malta has a strong two-party system, contrary to PR’s reputation for breeding third parties. There has been no serious third-party competition in Malta since 1966. Maltese politics are highly partisan and support is very evenly divided between the major parties. The narrow majorities common in Maltese elections illustrate STV’s ability to hand an election victory to the party with fewer votes, just as the plurality system can. After this happened in 1981, Malta amended its constitution to provide that if a party with a majority of the popular vote fails to win a majority of seats, it will have its numbers supplemented by enough additional seats to give it a parliamentary majority.

New Zealand used the first-past-the-post system for 140 years, with a detour into the second ballot in 1908 and 1911. In 1993, after two referendums, New Zealand adopted a “Mixed Member Proportional” system along German lines. Each elector casts a vote for a constituency MP and one for a party list. To sit in parliament, a party must secure one constituency seat or five per cent of the popular vote. Five seats are set aside for Maori MPs. The first MMP parliament had 65 constituency MPs (including the five reserved Maori seats) and 55 list MPs.

The new system came under fire after the 1996 election New Zealand politics seemed as acrimonious as ever, contrary to the expectations of the proponents of MMP. The behaviour of the New Zealand First Party, which spent two months behind closed doors deciding which party to join in a coalition, angered voters, as did the fact that parliament’s size had increased.

At the same time, the composition of the New Zealand parliament was more representative than at any time in history, with 30 per cent of MPs being women and seats being set aside for Maori members. There was evidence that voters took advantage of the tactical opportunity to split their ballots, and it seemed that they were not confused by the new system. But there was also evidence of disillusionment when the new system created “growing pains in the body politic.” Nevertheless, a delegation sent to New Zealand by the Jenkins Commission concluded that even if another referendum were held, New Zealanders would be unlikely to return to first-past-the-post, but would more likely modify the MMP system. New Zealand held its second MMP election in November 1999.

The Scottish Parliament, created by the Scotland Act 1998, uses a German-style additional-member system (AMS). Of the 129 Members of the Scottish Parliament (MSPs), 73 are elected in constituencies by the first-past-the-post method on one ballot. The other 56 are drawn from regional lists (each region comprising several constituencies) in proportion to the parties’ respective votes on another ballot. The first elections were held in May 1999, returning MSPs from seven parties (three with one MSP each).Iain McLean suggests that the introduction of PR into the Scottish parliament was eased by the fact that it was a newly created assembly; the forces of conservatism and inertia that might have played the spoiler in a change to PR in an established legislature were not present.

Britain is usually thought of as the bastion of the first-past-the-post system, but in fact FPTP as we know it did not become the dominant British electoral system until the Third Reform Act of 1884–1885. In 1917, Britain came close to introducing a combined AV/STV system for the House of Commons. In 1976 the Hansard Commission on Electoral Reform recommended an MMP system. Presently, aside from the Scottish (and Welsh) AMS experiments and the STV system in Northern Ireland, the United Kingdom uses a regional list PR system for elections to the European Parliament. More radical, though, is the proposal of the Independent Commission on the Voting System (the Jenkins Commission), appointed by the Labour government in 1997. Surveying the electoral landscape of the Westminster Parliament, the Commission concluded that Britain would be better served by “a two-vote mixed system which can be described as either limited AMS or AV top-up. The majority of MPs (80 to 85 per cent) would continue to be elected on an individual constituency basis, with the remainder elected on a corrective top-up basis .... Within this mixed system the constituency members should be elected by the Alternative Vote.”⁸

Some Scenarios for Electoral Reform

What follows is only a set of suggestions. Neither the proposals nor the mathematics that illustrate them represent the precise shape of a revised electoral system, but only a rough outline. This proposal seeks to make the case for a modification of the Prince Edward Island electoral system to include an element of proportional representation. It does not claim to be a mathematically exact portrayal of how such a system would function, or would have functioned in the past.

The disproportional results of recent Prince Edward Island elections provide a strong rationale for electoral reform. Taking the 2000 general election as an example, the Progressive Conservatives’ 57.9 per cent of the popular vote entitled the party to a healthy majority, but the first-past-the-post system outdid itself and awarded the winner 97 per cent of the seats. The two opposition parties, supported by 42.1 per cent of the voters, were left with a single seat between them.

A purely proportional result would have given the PCs 16 seats and a comfortable five-seat majority, the Liberals 9 seats, and the New Democrats 2. This outcome would have provided a healthier Legislature both for the governors and the governed. The opposition’s numbers would have been more than negligible, and the distribution of seats in the Assembly would have followed the popular vote. Such facts provide a good basis from which to argue that the electoral system needs modification.

Designing a new electoral system, or altering an existing one, requires us, first of all, to ask what we want elections to accomplish. If we require  majority governments at any cost, the current system should work admirably (in the PEI context, that is). But if we seek a system that reflects the choices of the voters, rather than using ballots as vague suggestions, it is worth considering adding at least an element of proportionality. This is the conclusion for which this paper argues. That said, such a reform need not be radical; in the relatively traditional and conservative context of PEI society and political culture, radical reforms are likely to be rejected out of hand.

A system of pure proportional representation, à la Israel, the Netherlands, or Ireland, then, is unlikely to be acceptable to most Islanders. Due to its  high district magnitude list PR would require a small number of ridings. At the most extreme, this could mean turning the entire Island into a single electoral district. More likely, there would be four or five districts, with five or six members each. These might be aligned with county lines or the boundaries of the four federal electoral districts. But this would mean eliminating the single-member constituencies, and for that reason it is probably unacceptable. No proposal that would dispense entirely with the representation of districts by individual MLAs is likely to be taken seriously. Even the most radical reforms proposed to the Election Act and Electoral Boundaries Commission did not contemplate eliminating or reducing geographical districts in this manner. The same consideration faced Britain’s Jenkins Commission, whose mandate was to devise an electoral system that would be more proportional than first-past-the-post without severing the MP-constituency link. For this reason above all, a pure list system seems inappropriate for PEI.

The single transferable vote presents its own set of problems. It, too, requires a district magnitude of three, four, five or more. Most Irish STV ridings have three or four seats; Maltese and Tasmanian districts have five members each. PEI could sustain no more than five or six five-member STV districts. This looks like another unacceptable deviation from the one-MLA, one-riding principle. STV is also burdened by a complicated vote- counting procedure, and the War and Peace of electoral ballots. If three parties run candidates in a five-seat district, the voter is presented with up to fifteen names to rank. On the other hand, STV provides a measure of flexibility difficult to achieve in a list system, since it permits voters to choose among different candidates of the same party, rather than voting for a list (although it should be noted that there are “open-ballot” list systems that allow voters to change the order of names on a party’s list). Ultimately, though, STV’s high district magnitude and complexity make it an unattractive prospect for PEI.

Putting aside the party list and Single Transferable Vote, the most promising choice for a new system is a mixed-member system roughly based on the models used in Germany, New Zealand, Scotland and Wales.

Such a system might be constructed by dividing the Legislative Assembly between a reduced number of single-member constituency seats and a remainder of proportionally apportioned seats, drawn from party lists, which might be considered “top-up” seats. The list seats would remedy, at least in part, the disproportional results of the constituency contests. For example you could have a Legislature of 30 members, of which 20 would be elected in single-member constituencies and 10 elected from party lists according to the parties’ share of the popular vote. Several presenters to the Election Act and Electoral Boundaries Commission advocated variations on this approach.

A mixed system, preserving the single-member constituency but adding a dose of proportionality to ensure a respectably-sized opposition, is the most likely option to be accepted on PEI.

There are still technical issues to resolve, such as whether to use one ballot or two. If two ballots are used, one would be for a constituency candidate and one for a party list. A single ballot would have the advantage of not requiring a change in procedure. Everyone would vote for a constituency representative and the popular vote would be calculated, probably on an Island-wide basis, from the vote across all the ridings. The list seats would then be distributed according to the popular vote across the province.

It would also be necessary to establish an electoral threshold, the minimum percentage of the popular vote necessary for a party to be entitled to take its proportionally apportioned seats. In pure list systems, thresholds range from less than one per cent in the Netherlands to 10 per cent in the Seychelles. In the MMP systems of Germany and New Zealand, the threshold is 5 per cent, with a “back door” whereby a party gets proportional representation if it elects a given number of constituency members (three in Germany, one in New Zealand). A threshold of between five and ten per cent might be appropriate for Prince Edward Island; the third party now has about 8 per cent support, and there is no indication of more parties appearing. Given PEI’s small size and general lack of internal cleavages, the “back-door” could be dispensed with. Its usual function is to give regionally based parties (as in the former East Germany) a fair chance to sit in parliament.

By making a series of assumptions, we can roughly estimate how such a mixed system would have operated in the 2000 provincial election. We will assume, for simplicity’s sake, that the Legislative Assembly was enlarged to 30 seats, 20 of them single-member constituency seats and 10 drawn from province-wide party lists. We will also assume that the threshold for proportional representation is set at 8 per cent (about the level of NDP support in the last two provincial elections). If the percentages of constituency seats won by each party remained as they were in the actual election, the PCs’ 97 per cent would translate into 19 of the 20 constituency seats; with one for the Liberals. As for the party list seats, the PCs’ 57.9 per cent of the popular vote would probably entitle them to six seats, while the Liberals’ 33.7 per cent would give them three, and the NDP would pick up the final seat on the strength of its eight per cent of the vote. In total, we can postulate an Assembly composed of 25 Progressive Conservatives (83.3 per cent), four Liberals (13.3 per cent), and one New Democrat (3.3 per cent).

While still disproportional, this is a much more balanced result than the plurality system provided. Furthermore, in designing the system it would be possible to adjust the degree of proportionality by increasing or decreasing the number of list seats and redistributing the single-member constituencies accordingly. More list seats mean more proportionality. For instance, if half the seats (15 of 30) were distributed by proportional representation, the PCs would win 9 list seats and 14 constituency seats, for a total of 23 (76.7 per cent); the Liberals would hold one constituency and five list seats, for a total of six (20 per cent); and the New Democrats would win a single list seat (3.3 per cent).

We might also consider emulating the German system more closely. A point that is sometimes overlooked in discussions of the MMP system is that, despite its “mixed” nature, it actually functions with nearly perfect proportionality. This is because the votes for party lists actually determine the number of seats to which each party is entitled. The list members simply make up the difference between that total and the number of constituency seats each party wins. If a party wins more constituency seats than its popular vote would entitle it to, it keeps the extra seats, and the legislature is temporarily enlarged. Applied to PEI, and assuming that there were 15 list seats and 15 constituencies, the most recent election would have had roughly the following result: the PCs’ 57.9 per cent of the vote would entitle them to 17 out of 30 seats. If they had already won 97 per cent of the constituency seats (i.e., 14 of 15), the extra 3 seats would come from the list. The Liberals 33.7 per cent of the popular vote would entitle them to 10 seats; if they held one constituency seat, this would mean they would name nine MLAs off their list. Finally, the New Democrats’ 8.4 per cent would give them 3 list seats. Of the three options considered here, this is by far the most proportional.

It is clear that an element of proportionality – even as a relatively small portion of the total number of seats – could ensure a more representative balance of parties in the Legislative Assembly, and enough opposition members to prevent a continuation of the pattern of virtual single-party legislatures seen over the past decade.

Editor’s Note: This study was presented to PEI’s Special Legislative Committee on the Election Act established in June 2000 and chaired by James Bagnall, MLA. The Committee held hearings and tabled its final report on April 24, 2001. It made a number of recommendations related to the administration of the Election Act. The Committee concluded there were too many unanswered questions to recommend implementation of a system of PR. Accordingly it called on Elections PEI to begin a review of systems of proportional representation in existence in other jurisdictions, paying particular attention to jurisdictions of comparable size. After the report has been received, it said Islanders should be broadly consulted on a specific system or systems. In December 2001 Elections PEI produced a Report on Proportional Representation. It included three possible scenarios but did not make any recommendation except that any “binding decision for one system over another system should be left to a provincial referendum, preceded by an impartial campaign of public education about the issues involved in he choice.”

Notes

1. Jean Halliday MacKay, The Home Place: Life in Rural Prince Edward Island in the 1920s and 30s (Charlottetown: The Acorn Press, 1999), p. 128. See also Wayne E. MacKinnon, “Island Politics and Government,” Harry Baglole, ed., Exploring Island History: A Guide to the Historical Resources of Prince Edward Island (Belfast, PEI: Regweed Press, 1977), 68-69; Clark, Franchise, 2-3.

2. Clark, Politics, 290; Frank MacKinnon, “Big Engine, Little Body, “ Martin Robin, ed., Canadian Provincial Politics: The Party Systems of the Ten Provinces (Scarborough, Ont.: Prentice-Hall of Canada, 1978), pp. 222-247.

3. Report of the Independent Commission on the Voting System (London: The Stationary Office, 1998):
http://www.official-documents.co.uk/document/cm40/4090/4090.htm, para. 47.

4. Vernon Bogdanor, What is Proportional Representation? A Guide to the Issues (Oxford: Martin Robertson, 1984), p. 195.

5. Douglas W. Rae, The Political Consequences of Electoral Laws (New Haven and London: Yale University Press, 1967; rev. 1971), pp. 15-39.

6. J. Denis Derbyshire and Ian Derbyshire, Political Systems of the World (New York: St. Martin’s Press, 1996 pp. 76 and 77.

7. Henry Milner. “Obstacles to Electoral Reform in Canada,” The American Review of Canadian Studies 24:1 (Spring 1994). p. 42.

8. Jenkins Report, Recommendations 1 and 2.