Changing the Canadian Electoral System

Louis Massicotte February 1st, 2001

This paper argues that changing the Canadian electoral system by adopting some form of proportional representation (PR) would have more advantages than drawbacks, but concludes that such a change appears unlikely in the near future.

The most common criticism of the existing system is that it fosters national division. Regional variations in party support are exaggerated by the electoral system, and the country appears more polarized than it really is. The system has also been criticized on other grounds. For example, it often grants power to a party supported by only a minority among the electorate, a scenario that an increasing number of Canadians find unacceptable. It has been accused of slowing the increase in women’s representation in the House of Commons and of driving down voter turnout. Some argue too that the electoral system and current party alignments together preclude any alternation in federal office for the foreseeable future.

Majority systems (the French double-ballot system and the Australian alternative vote, for example) will not solve these problems. Indeed, they might lead to an increase in electoral distortions. The two most credible reform options are (a) a full-fledged PR system whereby seats are distributed in proportion to the votes cast for parties in multi-member districts, or (b) a German-style mixed system where about half the members are elected in single-member districts, while the other seats are distributed in proportion to votes obtained.

Should either option be adopted, Parliament would become more representative and might include more women. Regional polarization would be reduced, with parties making inroads in regions of traditional weakness, although there is little evidence that this alone would solve the crisis of Canadian federalism. There would be less emphasis on constituency work for MPs if straight PR is introduced, though this would be mitigated under the German option.

Proportional representation would alter the Canadian pattern of governance. Coalition government or single-party minority governments would become the rule. Cabinets would be less durable, and the position of the prime minister in cabinet would be weakened, especially with regards to appointments to the Senate and the judiciary. Elections would be followed by negotiations between parties to determine the government’s composition.

Based on comparative research, the paper challenges the conventional wisdom that equates stable single-party cabinets with good governance. While in the past much has been made of extreme cases like Israel and Italy, more recent research strongly suggests that, based on indicators like GNP growth, unemployment and the incidence of riots and political violence, PR countries as a whole are no worse governed than countries with plurality or majority systems.

When changes to Canada’s political institutions are contemplated, whether by academics, pundits, or political actors, electoral reform is often near the top of the list. From among the manifold ways of translating votes into seats, Canada uses one of the oldest and simplest, first-past-the-post, wherein members are elected in single member constituencies based on a simple plurality of the vote. In the 1993, 1997 and 2000 federal elections, this system produced sizable distortions in the translation of votes into seats, especially for parties whose support was thinly spread across the country.

The influence of first-past-the-post on the longterm evolution of the federal party system was chronicled in the first Choices paper in the Strengthening Canadian Democracy series (Richard Johnston, “Canadian Elections at the Millennium”) as were some of its baneful effects on Canadian governance in the five-party era. In this second paper in the series, Louis Massicotte undertakes a critical examination of alternative electoral systems involving greater measures of proportional representation (PR). While recognizing some of the drawbacks involved in altering electoral systems, Massicotte concludes that the overall balance sheet comes up positive. Aside from the obvious benefits of PR — a fairer relationship between votes and seats and enhanced regional representation for most parties — he also considers other potential consequences, such as modified job descriptions for MPs and a reduction in prime ministerial power. The analysis throughout is couched in comparative context by invoking the experience of other countries that have long had PR systems in place.

In keeping with the overarching objective of the Strengthening Canadian Democracy project — to generate ideas for institutional reform that have some realistic prospect of implementation — we asked Massicotte to consider ways that electoral reform might actually come about. He identifies three possible avenues: the election of a minority government dependent for support on parties that stand to gain from electoral reform, a referendum allowing the Canadian people to vote on the issue, and a court challenge based on the Charter. Massicotte considers the first and third more likely routes, since referenda in Canada can only be initiatied by political elites, who are unlikely to do so if wary of the probable outcome. But there are significant hurdles under the other scenarios too. If the case for reform has intrinsic merit, parlaying that into momentum for change remains the biggest challenge.

Note: Professor Massicotte’s text was prepared prior to the November 2000 federal election. The results of that election and its implications for electoral reform are discussed in a short postscript.

Introduction

Canada belongs to the minority among the established democracies (together with Britain and the United States) that have kept the first-past-the-post electoral system used in the nineteenth century. Most of the other constitutional polities of Europe that had a plurality or a majority (two-ballot) electoral system at that time later switched to proportional representation (PR) or to a mixed system.

In these countries, plurality or majority formulas were criticized primarily because they resulted in national parliamentary outcomes that did not closely reflect the votes cast for each party. In our federally-minded country, the debate over the electoral system is predictably dominated by a different concern: preserving the cohesion of the federation by ensuring that federal political parties, especially the governing party, include elected members from all regions.1

This is not, however, the only argument for electoral system reform. Some have been concerned with the failure of the existing system to provide parties with representation corresponding to their electoral support in the country as a whole, while in recent years others have advocated PR as a device that would increase the representation of women in Parliament.2 Recently it has been argued that given existing party alignments, first-past- the-post condemns Canadians to decades under Liberal governments for lack of alternatives, and is driving down electoral turnout.

Some form of proportional representation has usually been advocated as a remedy to these problems, though other formulas like alternative voting or second ballots have also been advocated, if less frequently. This paper takes up the debate, assessing the merits of different electoral systems in the Canadian context and analyzing the current prospects for reform. We begin with a review of the chief arguments raised against the existing electoral system.

Problems with Canada’s Current Electoral System

Regional polarization

Regionalism has been a central feature of Canadian federal elections at various points in the country’s history. This has taken two forms. First, building on the perception that a region was disadvantaged by policy decisions of the federal government, political parties have emerged with a strictly regional agenda and electoral support in one region only. Second, even political parties with a more national perspective have received quite unequal support in the country’s regions, at times sweeping some while being shut out in others.

Some degree of electoral regionalism can be found in most democracies, whether unitary or federal, either because a few regionalist parties have been represented in parliament for a long time, or because some regions tend to support a specific party more than do others. Yet the extent of electoral regionalism, measured on the basis of seats won by parties, is higher in Canada than in Germany, Australia, Switzerland, Spain and even Belgium.

In the 1997 general election, all but two of Ontario’s 103 seats went to the Liberals. Elsewhere in the country the Liberals won the most seats in only three other, small provinces, namely Newfoundland, Prince Edward Island and Manitoba and in the Northwest Territories. Elsewhere, the Reform Party won a majority of seats in British Columbia, Alberta and Saskatchewan; the Bloc Québécois in Quebec; the New Democratic party in Nova Scotia and the Yukon; while the Progressive Conservatives won more seats than the other parties in New Brunswick. Thus, each of the five parties represented in the House of Commons came first in at least one province or territory. Further, the sectional nature of two of these parties was highlighted by the fact that the Bloc Québécois fielded no candidates outside Quebec, while the Reform party fielded only 11 candidates in Quebec’s 75 seats (in 1993, it had none) and was not able to gather a full slate of candidates in Prince Edward Island (2 out of 4), Nova Scotia (9 out of 11), New Brunswick (8 out of 10) or even Ontario (where one riding had no Reform candidate).

Regionalism was also manifest in both the 1979 and 1980 elections. There were no regional parties at that time, but the Liberals were practically obliterated in the Western provinces while the Tories failed to win more than a handful of seats in Quebec. Even before that, the tendency of voters in each region to back different parties had, at times, been a source of alarm for commentators. Writing about the 1921 election, for example, an historian commented: “the political map of Canada had been transformed into a Balkan nightmare … a jigsaw of isolated tribes, merchant guilds, and husbandsmen.”3 In 1917, the country had emerged bitterly divided from an election focused on conscription, with Laurier’s Liberals gaining 62 of Quebec’s 65 seats but only 20 of the 170 seats in all other provinces.

To a large extent, electoral regionalism is an undeniable fact of Canadian life and cannot be eradicated simply by modifying electoral rules. However, the first-past-the-post system has exaggerated this regionalism by amplifying both the strengths and weaknesses of parties in different regions. For example, the Liberals’ quasi-monopoly over Ontario seats in 1997 is based on less than half of the popular vote (49.5 percent), while they have won few seats in the four Western provinces in recent elections with 24 percent to 34 percent of the vote. Reform’s virtual sweep of Alberta’s seats in 1997 is based on 54.6 percent of the vote, while their respectable 19 percent score in Ontario brought no seats at all. In Quebec, the strength of the Bloc Québécois (44 of 75 seats, against 26 for the Liberals), conceals the fact that the two parties were almost evenly balanced in the popular vote (37.9 percent vs. 36.7 percent).

The existing electoral system has not created regionalism. However, in recent years it has rewarded parties with a strong regional appeal, and disadvantaged weaker, nationally-oriented parties that attract votes more evenly from one region to another. With almost as many votes as Reform, the Progressive Conservatives won many fewer seats in the two most recent elections because their vote was more evenly spread among regions.

An immediate consequence of this situation is that forming a regionally representative cabinet becomes more difficult, insofar as the governing party has few or no elected members in some provinces. Party caucuses are regionally skewed, and the policies adopted by governments may discriminate against unrepresented regions, or at least are vulnerable to such accusations. This process is somewhat self-perpetuating, since caucuses dominated by members from specific regions may slow down policy changes designed to widen support for the party in other regions.

Fairness

Fear of regional polarization and of its adverse consequences on Canadian unity has remained the driving concern among reformers. Yet Canadian election results are also vulnerable to the standard criticisms routinely raised elsewhere. Since 1921, federal elections have produced majority governments 15 times out of 23, yet in only three cases (1940, 1958 and 1984) has this been based on a majority of the popular vote. The governing Liberals’ majority at the 1997 election rested on a mere 38 percent of the national vote. Further, the electoral system at times has altered the rank-ordering of party strength. In 1993, the Bloc Québécois reached Official Opposition status while coming fourth in the popular vote. In 1979, Joe Clark formed a minority government just shy of majority status, yet his party had scored only 36 percent of the vote against 40 percent for the Liberals, whose popular support was heavily concentrated in Quebec. Results like these highlight that the plurality system is not only unfair in amplifying the representation of the leading party, but also because it sometimes gives power to a party that does not even have the support of a plurality of the electorate.

There is some polling evidence that a growing number of Canadians feel it is unacceptable that a party can form a majority government without winning a majority of the vote. In a survey conducted this year, 49 percent considered such an outcome unacceptable, 29 percent found it acceptable, and 23 percent had no opinion. Ten years earlier, in a similar survey, the respective figures were 39 percent, 34 percent and 27 percent.4

First-past-the-post makes it more difficult for new parties to emerge given that they have little chance of success unless their support is spatially concentrated. Supporters of PR argue that political debate would be enriched by the presence in the House of Commons of new political forces like the Greens.

Plurality systems are also accused of driving down electoral turnout, insofar as the supporters of a minority party in specific districts, knowing in advance that their chances of winning are slim, may choose to abstain. By the same token, supporters of the party that clearly dominates in a district may also conclude that voting is a waste of time. There is empirical support for this argument, as turnout tends to be higher in countries that have proportional systems.5 The decrease in turnout at federal elections to 69 percent in 1993 and to 67 percent in 1997 — down from a postwar average of 75 percent — lends greater salience to this dimension of the issue.

Under-representation of women

Women, who represent more than one-half of the total population, currently constitute only onefifth of the members of the House of Commons. As the percentage of women sitting in legislatures elected by proportional representation tends on average to be higher, many scholars have identified single-member districts as an obstacle to female candidates. The rationale is that multi-member districts require parties to field lists of candidates, making it easier to ensure gender balance on the ticket (and possibly ethnic and linguistic balance too). With single-member districts, parties tend to select candidates that fit the prevailing stereotype of the politician as a middle-aged male.

Equal gender representation in the House of Commons could also be achieved by keeping firstpast- the-post but switching to dual member districts, each electing one male and one female member, with each voter casting one ballot for each seat. Though this would solve the problem of women’s under-representation, it has not been advocated so far for Canada as a whole. A scheme of this sort was recently put forward for territorial elections in Nunavut, but was rejected in a referendum held in May 1997. The main reasons for its rejection seem to be that it would have prevented women from eventually outnumbering men in the Assembly and would have forced women to stand against other women.

Democratic alternation

The Liberal party has been described as one of the most successful parties in the democratic world, having been in office for a total of 73 years since 1896. Years of rule by another party, in practice by the Conservatives, have been comparatively few and far-between. There is no evidence that the first-past-the-post system is or has been significantly biased in favour of the Liberal party as such. Liberal dominance in Parliament has reflected their lead in the popular vote, and any party commanding a similar lead would likely have reaped the same rewards.

Richard Johnston has argued that the eruption of both the Reform party and the Bloc Québécois in 1993, combined with the precipitous decline in Conservative and New Democratic fortunes, has created a context that precludes any alternation in office for the foreseeable future. The two strongest alternatives to the ruling Liberals have little chance of making gains outside their respective bailiwicks, and indeed the Bloc has no intention of expanding beyond Quebec. While they are more nationally-oriented, the Tories are hampered by the impact of the Mulroney legacy. Should the Tory party vanish completely, polling evidence suggests that its supporters, in defiance of the view that the Tories and Reform are ideologically close, would mostly go to the Liberals.6 The result is that there is virtually no alternative to the Liberal party, each of the challengers combining enough negatives to be seen as unelectable. Johnston interprets the downward trend of voting turnout in the 1990s as a symptom that electors are exiting from the system because the alternatives seem unpalatable.7

The Case Against Electoral Reform

Most political scientists who have taken a stand on the electoral system have come down on the side of reform. Yet there is no unanimity, and various points of the case for electoral reform have been challenged.

A peripheral issue

A preliminary consideration is that at present few Canadians seem to care deeply about electoral reform. Issues like the health system or the Constitution arouse much more interest. The 1979 and 1980 election results generated a flurry of comments in favour of the introduction of some dose of proportional representation. In contrast, the contributions made by advocates of change following the 1997 election have not generated much reaction among politicians, columnists and the general public. While all four opposition parties represented in the Commons have expressed concerns about the existing system and shown interest in change, none is formally committed to a specific proposal, and Parliament’s attention in this area has been focused during the previous decade on electoral redistribution and on changes to the electoral machinery. There may be a malaise, but there is certainly no strong and widespread disgust with the existing system. Polls reveal some preference for PR among Canadians, but also disclose that they prefer majority over minority governments.8

No panacea for regionalism

The argument that first-past-the-post increases regional cleavages in Parliament is empirically strong, but the inference that a reformed electoral system would solve the problem has not gone unchallenged. Some critics have argued that the problem lies with the depth of regional tensions and the strength of regional parties more than with the electoral system. As Richard Katz has put it, “the fundamental problem is not how voter preferences are translated into a seat distribution in the House of Commons, but the distribution of the preferences themselves.”9 Others have pointed out that regional polarization comes and goes: the image of a country sharply and increasingly divided on regional lines that was suggested by the 1979 and 1980 elections was swept away in 1984 when Brian Mulroney won a majority of votes and seats in every province and territory. The acute electoral regionalism revealed by the 1921 election had similarly vanished after a few elections with the gradual demise of the Progressives.

It is also argued that Canadian prime ministers have dealt with the problem by appointing to the cabinet senators from unrepresented regions, and that the predominance of a specific region within the government caucus does not necessarily ensure that this region will automatically carry the day on every issue. As John Courtney has noted, even strong cabinet representation does not preclude the rise of regional alienation, as exemplified by the growth of Reform at a time when the cabinet included strong ministers from the West, or by the rise of the Bloc Québécois (and of separatism) under prime ministers from Quebec.

Problems with the alternatives

Opponents of electoral reform have also targeted the alternative formulas put forward. Proportional representation brings fair representation to political parties, but also generates many undesired consequences: political fragmentation, coalition or minority governments, cabinet instability, and changes of government unsanctioned by the electorate. It facilitates the emergence of new parties, but some of these could be extremist or farcical. Mixed systems also necessitate the coexistence in the House of locally-elected and regionally-elected members: might not the latter be branded as second-class MPs, or, on the contrary, as the privileged few who have little constituency work to do and owe their position to their high standing with the party brass?

Questionable effects on voter turnout

The argument that first-past-the-post drives down turnout and ensures the perpetuation in office of the Liberals is not flawless. Even granting that the average turnout in Canadian federal elections is lower than in most democracies and that the plurality system may be partly responsible, two points can be made. First, it is far from obvious that the decline of turnout in recent years has much to do with the fact that electors see no alternative to a Liberal government. The sharpest decline in turnout (six points) took place at the 1993 election, before the present pattern emerged. The decline recorded in 1997 (two points) seems comparatively mild. In addition, turnout has declined in most democracies over the last two decades, including those where PR prevails. Declining turnouts can more plausibly be traced to flagging trust in politicians and institutions among the Canadian public, a worldwide trend which became acute in the early 1990s.

No greater alternation in government

Finally, there is little evidence that PR guarantees the alternation of parties in office. Indeed, it arguably does the opposite by making it easier for some parties to rise, only to be shunned by others as unacceptable coalition partners because of their alleged extremism. Nor does it encourage opposition parties to fuse or to conclude alliances in order to offer a credible alternative. This can mean that the same group of centrist parties stays in power for a long time. There are plenty of examples of democratic PR countries that have been governed without interruption by the same party or coalition of parties for decades.10

Alternatives to the Present System

One of the chief problems faced by reformers is that Canadians are not very familiar with other electoral systems. Plurality rule has prevailed continuously at both the provincial and federal levels since the beginning of elections in Canada, except in three Western provinces, where alternative voting and/or proportional representation through the single transferable vote were tried and later discarded, the last one in 1956.11 The foreign country with which Canadians are most familiar, the United States, also uses a plurality system. In addition, alternative electoral systems are more complex than first-past-the-post, a feature that makes them more difficult to sell.

There are many types of electoral systems, though they fall broadly into four categories: simple plurality or first-past-the-post, as in Britain and Canada; majority systems, i.e. systems demanding more than one-half of the popular vote in a district in order to be elected; proportional representation; and mixed systems that combine proportional representation with either a plurality or majority system.

Majority systems

Majority systems are attractive for those who wish to keep single-member districts but want to raise the threshold for a member to be elected. As under first-past-the-post, in majority systems a candidate may be elected with a very small proportion of the vote, provided the remainder of the vote is divided between the other candidates, all scoring less individually than the leading candidate.12

There are two basic types of majority systems: alternative voting and the two-ballot system. The latter, used in France, is the more prevalent of the two internationally. In this system, the candidate, at the first ballot, needs a majority (more than 50 percent) of the popular vote in order to be elected. If no candidate reaches that threshold, a second and final ballot is held one or two weeks later. Then, two different rules may apply: either the candidate winning a plurality of the vote (as in first-past-the-post) is elected, or the competition is restricted to the two leading candidates on the first ballot, one of whom will necessarily be elected (this is known as a “run-off”). The first rule prevails in French legislative elections, while the second applies in most countries having a directly elected president.13

In the 1997 Canadian election, slightly more than one-third of the seats were won with more than 50 percent of the vote.14 If a majority had been required for election, a second ballot would have been necessary in almost 200 ridings. Much would have then depended on the second preferences of voters and on alliances between parties.

Second ballots, if they are to be held in many districts, entail a substantial increase in the cost of elections, at least for polling day. The alternative (or preferential) voting system, guarantees that a candidate will be elected with a majority of the vote, and with less expense. Voters are required to rank candidates on the ballot paper by marking “1” beside the name of the candidate they prefer, “2” beside their second choice and so on. On election night, only first preferences are counted. Candidates obtaining more than 50 percent of first preferences are elected, and other preferences are not examined. Should no candidate reach that threshold, however, the weakest candidate is eliminated and the second preferences marked on his or her ballots are transferred to the remaining candidates. This process is continued until a candidate reaches a majority of the vote. Alternative voting has been in force in Australia since the end of World War I. Though few other countries now use the system, it was in force for some time in three Western Canadian provinces for provincial elections.15

Alternative voting guarantees that winners will be backed by a majority of the vote in their respective districts, as does the two-ballot system provided the second ballot is a run-off between the two leading candidates. Otherwise, it does little to reduce the vote-seat distortions found in plurality systems. The chief advantage is to allow ideologically close but formally distinct parties to escape obliteration by combining their vote behind a single candidate.

What would have been the outcome of the 1997 election if an alternative voting system had been used? That a counting of subsequent preferences would have been needed in almost two-thirds of seats suggests at first sight that the outcome might have been substantially different, and that the Liberals’ weak majority in the House would have been jeopardized, especially by an exchange of preferences between Reformers and Tories in Ontario. However, a simulation conducted recently found otherwise.16 Projecting second preferences on the basis of the second choices of voters (using data from the Canadian Election Study), Bilodeau found that these would have altered the result in 31 ridings out of 301. The number of Liberal seats would have jumped from 155 to 173, while the number of Tory seats would have increased from 20 to 27. The representation of all other parties would have been smaller: 47 Reformers instead of 60, 36 Bloc Québécois members instead of 44 and 17 New Democrats instead of 21.17

The province-by-province results of that projection suggest that regional polarization would have been somewhat reduced, with the Liberals making substantial gains in Alberta and British Columbia (at the expense of Reform) and in Quebec (at the expense of the Bloc). However, neither Reform nor the Tories would have made any gains in Ontario, where the Liberals would have kept every one of their 101 seats. Further, the distribution of seats would not have been any “fairer,” that is proportional to the popular vote; with 38.5 percent of the vote, the Liberals’ share of seats would have increased from 51.5 percent to 57.5 percent.

Proportional representation systems

While the mechanics of PR are relatively complex, the basics are relatively simple. PR requires electoral districts with more than one member, preferably at least four. In each district, the number of votes cast for each party is divided by a quota, so that the “cost” of a seat is about the same for each party.

PR systems fall into two categories: list systems (by far the most common form) and single transferable vote systems (STV). The latter, conceived by Thomas Hare and advocated by John Stuart Mill, has been ardently promoted in Englishspeaking countries, so far with little success, but is the only variant of PR that has been used on occasion in Canada.18

The accuracy of representation depends mostly on the number of seats in each district (the higher the number, the more accurate), but also on the specific technique used for distributing seats (the highest average technique tends to magnify distortions and provide a “bonus” to the stronger parties, while the Ste-Laguë and largest remainder techniques produce more accurate distributions). Some countries also provide for national seats to top up the district results, so as to increase proportionality.19

Proportional representation is normally associated with more fragmented legislatures, as it is easier for a small party to win seats, and with the absence of a majority for a single party. As such, it often leads to coalition governments or minority single-party governments.

Mixed systems

Mixed systems combine PR with a plurality or majority system. A few basic combinations can be identified.20 The two most common are superposition (or parallel) systems, now typified by Japan, and corrective systems. The latter are in place in Germany, New Zealand, Scotland and Wales, and were proposed for Britain by the Jenkins Report in 1998. Superposition systems combine plurality or majority in single-member districts with PR in larger multi-member districts. Thus there are two sets of legislators, local and regional. Such is also the case in corrective systems, the crucial difference being that PR seats are distributed among parties so that the total number of seats (local and regional) won by each party approximates its percentage of the popular vote. Corrective systems result in proportional outcomes, but this is achieved while still retaining single-member districts for about half of all legislators. This is why such systems are officially known as “personalized PR” in Germany, and as “mixed member proportional” (MMP) in New Zealand. In contrast, superposition systems do not result in outcomes that are significantly more proportional.21

A survey, conducted in the mid-1990s, of 166 countries with working parliaments found 59 using the plurality rule, 25 with a majority system (either two-ballot or alternative voting), 56 using PR and 25 using a mixed system.22 The number of countries using mixed systems has increased dramatically in the 1990s.

Among other electoral formulas, the majority system must be ruled out as an alternative by anyone trying to respond to the criticisms of first-past-the- post cited above. A clear lesson from comparative studies is that the outcomes produced by majority systems are no more proportional than those produced by first-past-the-post, and are sometimes even less so, as was amply confirmed by the projection of the results of the 1997 Canadian election using the alternative vote.

Mixed systems have been most often advocated in Canada as alternatives to the existing system, because they offer the advantage of adding some dose of PR while keeping the single-member districts familiar to Canadians. However, numerous and substantially different variants have been put forward.23

What Impact Would PR Have on the Canadian Political Process?

This section explores the likely consequences of proportional representation for crucial features of the Canadian political system. The focus is on two options — full-fledged PR or a German-style mixed system — because they alone have any chance of both producing fairer results and addressing problems of regional polarization, low voter turnout and the under-representation of women.

“Full-fledged PR” is envisaged here as an electoral system where seats are allocated to parties in proportion to their vote in 4 to 10-seat constituencies.24 A “German-style system” means a mixed electoral system where about half of MPs are elected in single-member districts; other seats are allocated so as to make the total distribution of seats proportional to the votes cast for each party. Both formulas would result in proportional outcomes, the chief difference being that a Germanstyle system would retain a sizable contingent of members from single-member districts.

PR would have a direct impact on the shape and size of electoral districts, the work of members of the House of Commons and the representation of political parties and minorities. But it would also likely alter the way cabinets are formed and operate, their composition, their relationship with both Houses of Parliament, and the position of the prime minister. The federal-provincial balance might be modified as well. We should make no mistake about it: together with responsible government and disciplined parties, the first-past-thepost system is one of the three most crucial variables that have shaped Canada’s Westminster system of governance, both federally and provincially, and replacing it with PR would likely have sweeping consequences.

Not all the potential consequences of PR can be anticipated. We can determine what the result of previous elections would have been, assuming voters would have voted the same way. Such exercises are helpful, but the conclusions that can be derived from them are somewhat fragile, insofar as we cannot know what the distribution of the popular vote at future elections will be. It is even more difficult to anticipate with certainty the impact of PR on Canada’s system of governance, which will depend on decisions made by a myriad of political actors.

In these areas, the experience of countries where PR has been operating for generations is relevant, though it should not be transposed slavishly. The real world of PR encompasses dozens of countries, from the inauspicious cases of Italy and Israel to the more reassuring examples provided by Germany and the Scandinavian countries. We should resist the temptation to systematically assume rosy scenarios, if only because the occasionally wild expectations of reformers have often been shattered by the actual operation of the systems they ardently advocated.25

A more representative Parliament

Parliament, meaning here the House of Commons, would be more representative, insofar as the number of seats won by political parties would more closely match their electoral support. The number of political parties represented in the House would probably be higher than it is now, not necessarily because PR would lead to the fragmentation of existing parties, but because the threshold of admission for new movements like the Greens would be lower. The range of political views represented in Parliament would be broader, and would include viewpoints that Canadians might find innovative and engaging as well as some they would find objectionable, should any of these gather significant support.

Existing minority parties whose electoral support is widely spread would no longer be disadvantaged compared with parties whose electoral support is equivalent in size, but concentrated in specific areas. This would benefit the Progressive Conservatives and New Democrats.26

Countries with PR systems normally have a higher proportion of women in legislatures than countries with plurality or majority systems, which has led many activists and scholars to advocate PR as a kind of affirmative action measure to guarantee the presence of more women legislators.27 It is untrue, as some have argued, that the first-past-the post system blocks women’s access to Parliament. The number of women in the House of Commons now hovers around 20 percent, a significant increase compared with the all-male House elected in the 1968 election. Significant variation in the number of women among PR legislatures sheds some doubt as to whether PR alone guarantees the presence of women legislators. This led the Jenkins commission to conclude in its 1998 report that the evidence linking PR to women MPs was “not overwhelmingly strong.”28 Much depends on the extent to which political parties give priority to the inclusion of women on their slates of candidates. In the short term, PR in Canada would likely lead to more women in Parliament, though the percentage is likely to increase in the future whichever electoral system is used.

Certain ethnic and cultural groups would lose the advantage they now arguably derive from their concentration in some smaller metropolitan ridings. On the other hand, they might have some of their own included on party PR lists.

Less regional polarization in Parliament

PR would ensure that party caucuses include some representatives from most major provinces, provided, of course, they secure some minimum electoral support there. More specifically, PR would break the Liberals’ present monopoly over Ontario seats, which was a crucial factor in their back-to-back majority victories in the 1990s.29 It would mean that the Canadian Alliance, while losing some seats in the West, would gain seats in Canada’s largest province and to some extent alter the image of the Reform party as a purely Western party.30 Barring a major ideological realignment, it is unlikely, if PR were introduced, that the Bloc would be interested in even running candidates outside Quebec, or that the Canadian Alliance would secure more than a handful of votes in Quebec.

PR would provide all parties but the Bloc Québécois with more regionally balanced caucuses. It would become more difficult for Ontario or Quebec MPs to secure the inflated one-third of the ruling party caucus that the Quebec Tories achieved in 1988, the one-half that the Quebec Liberals achieved in 1980, or the two-thirds that the Ontario Liberals now enjoy.

This, in turn, would encourage parties to develop policies more palatable to all regions rather than engaging in “Churchill strategies.” On election night, regional variations would be accurately reflected in the distribution of seats, rather than being exaggerated, as was the case in 1979, 1980, and 1997. This consideration was a major factor in the decision taken in Belgium in 1899 to introduce PR, and was successful for a long time in reducing the polarization that existed previously under a majority system between Flemish Catholic and Walloon Liberal or Socialist areas.31

PR would resolve the deep-seated problem of regional polarization in federal elections, a phenomenon the Pépin-Robarts Commission claimed was a harbinger of the break-up of federations.32 After the 1984 election, many thought regional polarization to be dead following Brian Mulroney’s sweeping victory with a majority of both votes and seats in each and every province and territory. Yet this appears to have been a short interlude, with polarization resurfacing later with a vengeance.

The results of the 1998 German election illustrate very well how regional polarization can be reduced by PR. The Social Democrats won almost 65 percent of single-member district seats (elected, as in Canada, under first-past-the post rules). Without the addition of corrective seats, the opposition Christian Democrats would have been shut out in no less than seven of Germany’s 16 Länder. In the second largest Land, Bavaria, the ruling party would have won a mere 15 percent of the seats. The use of personalized PR resulted instead in the Social Democrats winning 37 percent of Bavarian seats, while the Christian Democrats were able to win seats in every Land.

This is not to say that PR will solve the crisis of Canadian federalism. The strong presence of a region within the federal executive, bureaucracy and judiciary does not necessarily dampen aspirations for regional autonomy, as shown by the rise of separatism in Quebec over the past 30 years. Many regional champions expect far more than mere inclusion in federal decision-making circles. However, the presence of a sizable group of cabinet ministers from these regions is likely to weaken their case: we can only imagine how disenchanted with federalism Quebecers would have been if their presence in the federal cabinet had been as weak for one or two decades as it was for a brief time after the 1957 and 1979 election.

Less emphasis on constituency work for MPs

Would PR change the job description of MPs? Here, the two PR systems should be distinguished. Full-fledged PR would substitute large, multimember districts for smaller, single-member districts. German-style PR would mean — assuming the total number of seats in the Commons remained around 300 — that half the MPs, possibly more, would continue to be returned from single- member districts, but that these districts would be twice as large as existing ones. The PR members would likely be returned from province-wide districts, except in Ontario and Quebec where the creation of regional “top-up areas” would probably emerge as the most appropriate solution.

In both cases, the job description of MPs would be affected, but to different degrees. If all MPs were elected in larger multi-member districts, they would find it more difficult to control their district associations than they do now with smaller, single-member district associations. Campaigning might become more onerous, unless candidates from each party informally decided to focus on a portion of the district. Constituency work, an activity which absorbs much energy and enhances the self-esteem of many MPs, would be affected by PR. In a full-fledged PR system, members would no longer be the sole district MP, but would face competition, within a wider territory, of other MPs from their own or other parties. There is no guarantee that within a wider district all areas would be equally covered, since most or all MPs could end up residing in a major urban centre. In mixed rural-urban districts, residents of outlying areas might be less likely to have a member close at hand. This is not a major problem in most PR countries, because constituency casework is a less prevalent practice.

A German-style system would maintain the close relationship between MPs and their enlarged constituencies, but would create a second layer of representation. As such, it is open to the often made (but rarely substantiated) charge that mixed systems produce two warring “classes” of MPs. How the two categories would interact cannot be predicted with absolute certainty, but it is worth pointing out that in the two dozen countries with mixed systems, no tensions are reported.33 In New Zealand, however, PR members are perceived by some to be “second-class” MPs,34 though in practice there is little to substantiate that perception.

No less party discipline

Currently members are constrained by party ties, and some, together with many observers of Parliament, find party discipline too constraining and are pushing for a higher number of free votes. Would PR lead to a relaxation of party discipline?

Little change should be expected on that front. There is no clearcut correlation between party discipline and single-member district systems, as exemplified by the contrast between the US House of Representatives and the Canadian House of Commons. In the former, parties exhibit much less cohesion at congressional roll-calls than members of other legislatures, a pattern of behaviour that has prevailed for over a century.35 In Canada, the same electoral system now coincides with relatively tight party discipline, yet our parties were much less cohesive during the second half of the nineteenth century.36

There is no reason to believe that individual MPs would have much more freedom and clout if PR were introduced. Disciplined parties appear to be the norm in all democratic countries except the United States, irrespective of the electoral system used.37 Party discipline is no less stringent within Canadian governing parties in minority than in majority situations.38 Indeed, at times it has been more stringent in the former. The practice in PR countries is for interparty negotiations to take place at the cabinet level, with MPs from all sides expected to accept the outcomes reached by their leaders. If the electoral fortunes of candidates rested on their party-determined position on party lists, parties would arguably be in an even better position to exact conformity from their followers in parliament. Further, MPs breaking with their respective parties would have more difficulty getting re-elected as independents in substantially enlarged districts.

No more single-party majority governments

If the past is any guide, single-party majority governments like those we have had for most of our history would likely become exceptional. The experience of PR countries suggests that singleparty majority governments would become rare interludes in a long succession of minority and coalition governments. Indeed, in countries where coalition government is the norm, it is not infrequent for a party having secured a majority on its own to maintain its earlier alliance with smaller parties, in anticipation of a return to the standard pattern.

Though Canadian parties have some experience of single-party minority governments (8 of the 23 elections held since 1921 have produced such outcomes), the latter would probably not become the standard government formula. So far, they have been resorted to as a temporary expedient by parties, in the hope that the ensuing election would produce a majority. The latter scenario is plausible with the first-past-the-post system (three elections39 since 1921 resulted in a majority for the incumbent minority government), as gaining only a few percentage points of the popular vote may well be enough to reach a majority of seats. Under PR, this is unlikely.

The experience of PR countries also suggests that coalitions would be more frequent than minority governments. For the years 1945 to 1987, Laver and Schofield found an almost two-to-one ratio in European cabinets.40 In this regard, much would depend on the constitutional rules governing cabinet formation. It is easier for minority governments to be formed and to endure if no formal vote of investiture in parliament is required for a new cabinet to be appointed. By the same logic, mechanisms like that provided by s. 49 of the French Constitution, whereby cabinets can be censured by the Assembly only by an absolute majority of its membership, with the votes in favour of censure being the only ones to be counted (which means that abstentions are implicitly counted as supporting the cabinet), would facilitate the survival of minority administrations.

If coalitions became the norm, Canadians would find little guidance in their own parliamentary history as to how to operate them. Ottawa’s experience with coalitions is uninspiring, being limited to the Borden Unionist coalition of 1917-1920. Ontario had a coalition of United Farmers and Labour under Drury (1919- 1923). British Columbia was ruled by a coalition of Liberals and Conservatives from 1941 to 1952 under John Hart and afterwards Byron Johnson. Saskatchewan had a Tory-dominated “Cooperative” coalition during the Depression years under Anderson (1929-1934), while a Liberal joined the NDP Cabinet following the inconclusive 1999 election. Bracken’s long premiership in Manitoba included a lengthy period of coalition government. An encouraging feature is that most of these coalition governments lasted for the full life of a legislature.

Despite our own limited experience in this field, coalition government is well entrenched in most democratic countries, including some with majority systems like France and Australia, and Canadians could derive inspiration from the practices that have been developed elsewhere.41

Less durable Cabinets

Our experience with minority governments (as opposed to coalitions) strongly suggests that if these became the norm under a PR system, governments would be less durable. Since 1867, minority governments in Ottawa have lasted an average of less than 20 months, compared with more than 50 months for majority governments. There is no reason to believe that minority governments, if they were formed in a PR context, would be any more lasting.

As our experience with coalition governments in Ottawa is quite limited, we must turn to the experience of other countries to see whether coalition governments would survive longer. In their study, Laver and Schofield found single-party minority governments to have lasted an average of 19 months in office, compared with 33 for minimal winning coalitions and 45 for majority governments.42 There was no evidence that coalitions had become more durable over time.43 True, one can cite examples of cabinets in PR countries lasting for the full duration of a legislature, of heads of governments serving aggregate terms comparable to those served by many Canadian prime ministers,44 or even of coalitions lasting for decades (like the Swiss four-party coalition that has run the country for the last 40 years). Yet the possibility remains that party fragmentation, coupled with the presence in Parliament of parties deemed “extremist” and systematically shunned as coalition partners, can combine to produce a succession of short-lived governments. Cases like Israel, Italy, the French Fourth Republic and Finland, where governments have on average been short-lived,45 are not necessarily typical of outcomes under PR, but they are possible.

Coalitions are inherently more fragile than single- party majority governments and are more likely to break up during the life of a parliament or to lead to early elections. When an unpopular decision has to be made, it is tempting for the junior partner to withdraw support in the hope of escaping voters’ vengeance. When an unforeseen issue arises, coalition partners may find their respective positions irreconciliable and dissolve their partnership.

It is difficult to gauge how Canadians would react to this new pattern of parliamentary politics. Criticisms have been voiced in recent decades about governments having too much power, the executive dominating Parliament and the prime minister behaving like an elected monarch.46 PR would likely make governments more fragile, but this may be what many Canadians actually want, especially if it means governments are more willing to listen and compromise. It is striking that Australia, the country where the working of the Westminster model arouses the least opposition, is also the only one where the power of the ruling party or coalition is checked by a PR-elected second chamber rarely controlled by the government party or coalition.

Assuming the next elections return the same parties we now have, cabinet formation and survival would be made more complex by the presence of the Canadian Alliance and the Bloc Québécois. The latter was created in order to highlight the deficiencies of the Canadian federation, not to support, or be part of, the federal cabinet. In the present context, it seems doubtful the Bloc would be a willing coalition partner, or would be accepted as such by the other parties. The Canadian Alliance would probably be more eager to join a coalition. How other parties would react is uncertain. If both the Canadian Alliance and the Bloc, with approximately 30 percent of the seats in the House, were deemed unpalatable coalition partners by the other three, the range of government formulas would be limited and cabinets would be more fragile.

Weaker Prime Ministers

In the long run, PR would probably erode the authority of the prime minister within cabinet. At present, prime ministers enjoy a very strong position and are acknowledged to be far more than primus inter pares. Compared to their counterparts in Britain, Australia and New Zealand, Canadian prime ministers have stayed longer in office and have been immune (so far) to cabinet or caucus revolts,47 probably because their status as party leaders derives from elected delegates at a party convention or direct election by party members, rather than from a caucus decision. This predominance also results from their position as leaders of the sole ruling party and from the existence of crucial powers commonly acknowledged to be personal prerogatives of the prime minister, like recommending the convocation or dissolution of Parliament, and making appointments to the bureaucracy, the judiciary and the Senate.

These prerogatives would survive intact in minority single-party cabinets, though the more precarious position of the cabinet as a whole would affect their use. In coalition cabinets, many ministers would belong to a party other than the prime minister’s and would have more complex loyalties. The list of the prime minister’s personal prerogatives is then likely to diminish. In the long run junior coalition partners are unlikely to tolerate appointments to the bureaucracy, the Senate or the bench being made secretly by a single individual. They are likely to insist on some kind of sharing of order-in-council appointments. They might even have a veto on the prime minister’s appointment.48

Inasmuch as prime ministers wield too much power, PR would likely make our cabinet system more collegial and less monarchic in its operation and style.

The relationship between election results and government formation

It has been argued that first-past-the-post empowers the electorate to select rulers “directly,” since elections normally result in a clear majority for one party with a recognized leader and clear policy positions, instead of leading to negotiations between parties as to what kind of coalition will be formed, who will lead it and what that government will do.

There is much truth in this argument, though most of the time a “clearcut” outcome reflects the will of only a plurality of the electorate. However, malapportionment of electoral districts or excessive concentration of a party’s vote in some districts at times lead to majority governments that do not even rest on a plurality of the popular vote, as is the case now in Quebec and British Columbia. This unquestionably amounts to a serious distortion, some would say a denial, of the voters’ will.

Under PR, voters would have a less direct say in government formation than they now do. Governments would be formed after the election through negotiations between parties, taking into account, of course, each party’s respective strength. Leaders may state in advance of polling day with which party they would ally — or not ally — but there would be no legal obligation for them to do so. There does not appear to have been any systematic survey determining to what extent, in PR countries, parties make such pre-election commitments. In the absence of party statements on their coalition partners, the feeling may develop among the electorate that the people’s role at elections amounts merely to “redistributing the cards” between political elites, while the most crucial decision of forming a government is in practice transferred to the latter. An even worse scenario can be imagined, where pivotal parties choose to ally with other parties in defiance of their own pre-election public statements (as occurred in New Zealand in 1996) or even to switch sides in the middle of the life of a parliament (as the West German Liberals did in 1982).

An encouraging consideration is that in these two cases, the “slippery partner” suffered losses in the ensuing election, thus deterring many politicians who would be tempted to emulate this behaviour.

No evidence that governance would be worse

It is still largely accepted in Canada, especially among political elites, that the first-past-the-post system, while distorting to some extent the representation of parties in Parliament, should be maintained because the stable and effective cabinets it produces ultimately ensure better governance.

Most Canadian elections have resulted in majority governments. Parliaments with no single- party majority have been typically short interludes managed by a single-party administration biding its time before it could call another election in the hopes of securing a majority. Except in the 1920s and in the 1960s, there have been few back-to-back minority parliaments that might have entrenched minority governments or coalitions as standard government formulas. Since 1980, no election has failed to return a majority government.

In such cabinets, decisions can be made quickly. It is easier to reach consensus within the confines of a cabinet composed of people belong ing to a single party. Policies that are unpopular in the short-term but advantageous in the long run can be pursued. The conventional wisdom then is that our present system produces firm and decisive leadership. For many Canadians, this is to be equated with good governance.

Until the 1970s, this was the accepted wisdom among students of comparative government. The Westminster system was widely acknowledged to be the most successful variant of parliamentarianism. After all, it originated in one of the most powerful countries in the world, one where democracy had successfully withstood the challenges of the interwar period.

However, more recent literature casts serious doubts on the governance advantages that supposedly derive from single-party majority governments. Simple assumptions that were deemed self-evident and accepted without question, have recently been tested and found wanting. The argument advanced by perceptive observers of French politics before 1958 like André Siegfried and Raymond Aron, namely that cabinet instability, while perhaps exposing a country to ridicule, is less harmful for governance than many assume, has been re-stated. There is, it turns out, no statistical evidence that economic growth in majoritarian countries is higher, or that inflation and unemployment are lower. On the whole cabinets are more stable in majoritarian countries than in PR countries, but there is no evidence that cabinet durability results in better governance outputs.

Arend Lijphart, for one, has attacked the conventional wisdom that assumed the existence of a trade-off between accurate representation and good governance.49 Lijphart reaffirms that PR coincides with a more accurate representation of parties in the legislature, a higher proportion of women, and a higher electoral turnout. He does not deny that executive durability is higher in countries with majoritarian electoral systems. Rather, he presents evidence that PR countries do not perform less well than countries with plurality systems on a number of important indicators. On average, countries with plurality systems have a lower incidence of political riots, but a higher incidence of political deaths than PR countries. On crucial economic indicators like economic growth, inflation and unemployment, countries with majoritarian systems do not, on average, outperform PR countries. In other words, there is evidence that PR leads to less durable executives, but no evidence that durable executives produce better policies. Indeed, some data point to the opposite conclusion. “Majoritarian governments,” Lijphart writes, “may be able to make decisions faster than consensus governments, but fast decisions are not necessarily wise decisions.”50

Lijphart’s conclusions are an important milestone in the age-old debate between supporters of PR and advocates of majoritarian systems, a debate that had become largely repetitive by the 1980s. His findings are grounded on the analysis of as many as 36 stable democracies, small and large, while the conventional wisdom tended to focus excessively on a few large and dysfunctional PR countries. Methodologically, they are based on factual quantitative indicators rather than on impressionistic evidence. To date, no systematic rebuttal has been provided by supporters of the Westminster model.

Other recent works have also offered a more positive assessment of governance under nonmajority administrations. Kaare Strom has analyzed the workings of minority governments in 15 democratic countries in the period from 1945 to 1987. He found that minority governments (either coalitions or single-party governments) were frequent, accounting for almost 35 percent of all cabinets formed. Looking at the performance of minority governments, he concluded that “contrary to conventional wisdom, minority governments do not perform particularly poorly in office. While minority governments are less durable than majority coalitions, they fare better at the polls and resign under more favourable circumstances. [They] perform best in those political systems where they are most common, and least well where they are most rare.”51 Following a more detailed examination of minority governments in Italy and Norway, he concluded that “at least in these countries, minority governments are just as effective as majority coalitions.”52

The thrust of this analysis is that many Canadians’ instinctive preference for majority governments and distaste for coalitions or minority governments rest on shaky foundations. Even if cabinets were shorter-lived than they now are, it is far from certain that Canada’s governance would perforce be negatively affected.

The federal/provincial balance

Any examination of the consequences of introducing PR federally must take into account the fact that Canada is a federation with powerful provinces and a strong secessionist movement, where federalprovincial relations tend to be conducted in an adversarial mode. Indeed, in 1983, former Clerk of the Privy Council Gordon Robertson described Canada as “the most quarrelsome” among major federations, and this remains true today.

Would a succession of short-lived administrations in Ottawa, facing strong single-party majority administrations in the provinces, hamper the stature of the federal government? It is difficult to derive significant insights from the experience of other federations because the same electoral system tends to prevail at both levels in federal countries, either because such congruence is constitutionally mandated (as in Austria), or because the federal and state legislatures have freely opted for the same kind of system, as in Switzerland.53 This pattern is probably due to the assumption that electors will be overwhelmed by the existence of two systems using different rules and criteria to elect national and state legislators.

Introducing PR for elections to the House of Commons might have a coattails effect on the provinces. Assuming it did, there would nevertheless be an interval between both sets of changes during which more fragile federal cabinets would have to deal with stronger provincial executives. Except in Quebec,54 little consideration has been given in recent years to PR in the provinces, despite striking anomalies like governing parties being returned to office in Saskatchewan (1986), British Columbia (1996) and Quebec (1998) with a smaller proportion of the vote than the main opposition party.

It is not, however, necessarily the case that PR would result in less assertive federal cabinets in the field of federal-provincial relations. It is true that Ottawa’s most prolonged succession of minority administrations (1962-1968) coincided with a more accommodative attitude with the provinces, as exemplified by the pensions deal of 1964, the abortive Fulton-Favreau formula which granted every province a constitutional veto, or the establishment of the Royal Commission on Bilingualism and Biculturalism. But it is also true that Trudeau’s attitude toward provincial governments does not appear to have been markedly more accommodating while he was heading a minority cabinet from 1972 to 1974, as exemplified by his energy policy. Further, Mulroney’s very large majority in the Commons during his first term coincided with one of the most accommodative eras in federal-provincial relations, which included the mothballing of the National Energy Program and the negotiation of the Meech Lake Accord. One is tempted to conclude that in this field, much more depends on the character and beliefs of the federal prime minister than on the extent of his or her support in the Commons.

Even if federal cabinets were weaker, however, this might be counterbalanced to some extent by an increase in their representativeness. More broadly-based federal cabinets might weaken the claim of some premiers, when facing a federal cabinet including no minister from their own province, to be the true spokesperson for provincial interests. William Irvine, one of the leading advocates of PR in the late 1970s, argued that PR, by increasing the representativeness of the federal government, might spare Ottawa massive transfers of powers to the provinces.

Quebec and francophones

How would PR affect francophones in general and Quebec in particular? PR would not change the proportion of Quebec seats, which would remain governed by section 51 of the Constitution Act, 1867. However, it would alter the shape of political representation in that province and others. Bakvis and Macpherson have documented the fact that Quebec’s “block vote” made a substantial difference to the outcome of many Canadian elections throughout the 20th century.55 This resulted not only from the sheer weight of Quebec seats (between 25 and 29 percent of the total, depending on the election) but also from the fact that Quebecers tended historically to support massively a specific political party, thus enhancing the chances of that party forming the government. An extreme illustration of this occurred in 1980 when all but one of Quebec’s 75 seats went to the Liberals, allowing this party to form the government despite trailing in all other provinces combined. For the next four years, more than half of the ruling party caucus came from Quebec. While the Liberals’ share of the vote in Quebec was already high (68 percent), only the first-past-thepost system could translate it into 98.6 percent of the seats. In this sense, PR would reduce Quebec’s clout — or for that matter the clout of any major province engaging in block voting to the extent Quebec did until 1984.

Before rushing to the conclusion that PR would hamper Quebec, two notes of caution must be added. First, block voting is a risky game, as any region which puts all its eggs in the same basket may secure two kinds of results: either strong representation within the winning party (as Quebec usually got), or very weak representation, as in 1917, 1957 and 1979. In the latter scenario, a region will find itself with a limited number of cabinet seats, or must rely on expedients like ministers sitting in the Senate, which in retrospect does not seem to have been an effective substitute either for Quebec (1979-80) or for the Western provinces (1980-84). PR, on the other hand, would guarantee each major party (with the possible exception of the Canadian Alliance) a minimum number of seats from Quebec, and would substantially reduce the likelihood of the province being severely underrepresented in Cabinet.

Second, one might doubt whether massive oneparty contingents from Quebec will be frequently elected again. The Liberals’ historic dominance of federal elections in Quebec rested during the 1970s and the early 1980s on the decision of supporters of Quebec sovereignty to stay away from the federal arena. This factor appears to have contributed strongly to Trudeau’s lopsided victories, at a time when the PQ was doing well on the provincial scene but was unwilling to divert its energies to federal elections. Many PQ supporters abstained, deliberately spoiled their ballots, voted for fringe parties like the Rhinoceros Party, or dispersed their vote more or less strategically among the Conservatives, the Créditistes, the New Democrats, or even the Liberals.

Since 1990, supporters of Quebec sovereignty have had their own party on the federal scene, fully backed by their provincial allies. A return to Liberal ascendancy in Quebec and the election of massive Liberal contingents from that province appear unlikely unless the sovereignty option vanishes completely from the Quebec political scene. The most likely scenario for the near future is that the electorate in Quebec will remain fragmented and that there will continue to be a substantial number of Bloc Québécois members in the House of Commons.

PR would have been helpful to the Bloc Québécois had it existed in the 1970s, but would likely be detrimental to the Bloc now. First, it would eliminate the advantage the Bloc has derived at the previous two elections from the division of the federalist vote between the Liberals and the Progressive Conservatives. It would increase the likelihood of prominent Quebec federalists being elected in heavily francophone areas, thus undermining the Bloc’s claim to speak for Quebec’s francophones. PR would also diminish a handicap that afflicts the Liberals and benefits the Bloc, namely the excessive concentration of Liberal support in English-speaking areas of Quebec. This phenomenon, which is replicated to an even greater degree on the provincial scene, largely explains why the nearly equal vote shares for the Bloc and Liberals in the 1997 election (37.9 percent vs. 36.7 percent) were not matched in terms of seats (44 Bloc MPs vs. 26 Liberals). The massive majorities won — and thus votes wasted — by the Liberals in western Montreal provide the explanation.

On the other hand, PR might reduce the likelihood of francophones being elected from provinces other than Quebec. The reasoning here is that PR implies much larger electoral districts, and much would depend on the willingness of parties (not only national leaders, but also local activists) to include French-speaking candidates on top positions on their lists. For example, under its present boundaries and linguistic profile, Ottawa-Vanier can be expected to return a francophone MP, but the same result would be less certain in a larger constituency including the whole city of Ottawa and its satellite communities, where francophones make up about 10 percent of the population. If New Brunswick were a single PR constituency electing 10 members, Acadian representation could also be reduced, depending on decisions made by parties while preparing their lists of candidates. By this reasoning, a German-style mixed system, by keeping single-member seats, would help to preserve the representation of Acadians.

The role of the Senate

Two important constraints influence the impact of the Senate on the political process. The first, that Senators are appointed rather than directly elected, diminishes the clout of the Senate insofar as most Canadians believe it is illegitimate for a chamber so constituted to oppose the government on major issues. This would not be affected by PR. The second constraint depends on whether the Senate is dominated by the government or the opposition: experience suggests that the Senate is much more likely to be assertive when the government has no majority “up there,” as exemplified by the behaviour of Senators from 1984 to 1990 and to a lesser extent from 1994 to 1996. As long as the Senate continues to be made up almost exclusively of members from both traditional parties, coalition governments including those parties can expect a cooperative attitude from Senators. Coalitions excluding those two parties can expect a rough time in the Senate until they have appointed enough of their own in that chamber. It is likely, furthermore, that junior coalition partners will sooner or later insist on some share of Senate appointments, thus leading to a more broadly representative Senate than is presently the case.

The Governor General

Governors general are appointed and may be dismissed by the Queen at the personal request of the prime minister. Their political influence is extremely modest, notably because most elections produce clearcut results, and because even in minority contexts, refusing to follow the prime minister’s advice in the use of the reserve powers is assumed — rightly — to be highly risky. In coalition cabinets, junior partners might challenge the right of the prime minister to personally recommend appointments to Rideau Hall.

Would PR, by multiplying hung parliaments, provide an opportunity for the governor general to play a more active role in the selection of the prime minister, as parliamentary presidents have in some PR countries? Probably not. The trend in the latter, including New Zealand, is for party leaders to negotiate between themselves, and for the head of state to be informed of their conclusions and to act accordingly. Unlike hereditary monarchs or directly or indirectly elected presidents, the governor general may be dismissed by the Queen at any time on the recommendation of the prime minister. This should dampen any temptation by the governor general to act as a referee, either in cabinet formation or with regards to the dissolution of Parliament.

The judiciary

The relative position of the judiciary in the Canadian political system would not be directly affected by PR, as it derives from constitutional provisions. The only possible influence has to do with the appointment of judges. Junior coalition partners would probably insist on having a more decisive input in judicial appointments, or on a more open selection procedure involving public hearings by Commons committees.

No threat to the survival of democracy

Some still fear that PR would endanger the very survival of democracy. A quick succession of shortlived cabinets might give the public the impression that anarchy prevails at the highest level of the state, and the country could become a laughing- stock abroad. At worst, the working of parliamentary government under PR might be unfavourably contrasted by the public with the firm and decisive leadership of earlier monarchs or dictators, thus discrediting democracy itself.

The argument that PR can lead to democratic breakdown, quite popular in the 1940s and 1950s, is now discredited, though it may have some validity for emerging democracies.56 This argument was inspired by an analysis that focused excessively on the experiences of Weimar Germany and interwar Italy. It overlooked the fact that in many PR countries like Switzerland, the Low Countries and Scandinavia, democracy survived quite well the challenges of the interwar period. A recent survey of historical works on democratic breakdowns during the interwar period revealingly makes no mention of PR.57

Scenarios for the Introduction of PR

In what kind of circumstances can we foresee PR being introduced in Canada? Socalled “rational choice models” assume that PR will be accepted when the party or parties in power reach the conclusion that this system will work to their advantage. This has been the case, notably, of ruling parties that were anticipating oblivion at the next election under a plurality or a majority system. PR could also, however, be imposed on politicians by the voters through a referendum. Finally, one can imagine PR being forced on Parliament by court rulings.

The parties

None of the parties represented in the Commons has pressed very hard for reform of the electoral system. Nonetheless, all but the Liberals have shown some interest in the topic, as expressed by the adoption of resolutions at national conventions over the past three years.58

The Progressive Conservatives and the New Democrats agree that electoral system reform is an ”important issue.” Both these parties and the Canadian Alliance cite proportional representation as a possible solution. In addition, the Conservatives add a run-off system as a possible alternative while the Alliance mentions the single transferable ballot. The Bloc Québécois cautiously envisages the introduction of some measure of PR, though this clearly refers to the electoral system of an independent Quebec rather than of a reformed Canada.

Specific concerns expressed by both the NDP and the Conservatives about the existing system include its unfairness, to which New Democrats add regional polarization. The Conservatives are also concerned about the extent of prime ministerial power under the existing system. The Alliance is the only one that has promised to submit a new electoral system to voters in a referendum. None of these parties, however, has a platform as specific as the Green Party, which in 1998 endorsed a German-style mixed system. The Greens are also the only party to cite the increased representation of women and minorities as an advantage of their proposal.

Policy statements by opposition parties reveal at least some awareness that a problem exists and should be addressed one way or another. However, they should be taken for what they are: stands couched in fairly general terms that may or may not be implemented once in office. For example, the Parti Québécois has been committed to a reform of the provincial electoral system since 1970 and has been in office with a majority for a total of 15 years since then. Yet it has failed to deliver so far and has shown no intention of doing so since its return to power in 1994.

The introduction of PR by the existing Liberal administration appears unlikely at this point. Despite the sympathy Pierre Trudeau expressed for some dose of PR after his temporary retirement as Liberal leader in 1979, the Liberals have subsequently shown no interest in the matter, even during the period (1980-1984) when the paucity of their representation in the Western provinces was perceived as a serious problem. Indeed, Liberal backbenchers at that time even resisted the introduction of PR for an elected, though not very powerful, Senate. Presently, the trade-off PR would produce between a weaker Liberal representation in Ontario and a stronger Liberal representation elsewhere is unattractive for Ontario Liberals, as it seemed in the early 1980s for the then strong contingent of Quebec Liberals. Those Liberals whose chances of election would be enhanced by PR are simply not in Parliament to make their case, while those who would be adversely affected by it now dominate the Liberal caucus.

However, these circumstances might change. One can imagine, for example, the Liberals losing their majority in the Commons at a subsequent election while remaining the strongest party, and being forced to conclude an alliance with the Progressive Conservatives or the New Democrats. As these two parties now have everything to gain from electoral reform, either might try to exact the introduction of PR in exchange for its participation in a Liberal-led coalition. Whether the Liberals would accept a change of this magnitude is, to say the least, uncertain — because PR would entrench the stronger bargaining position of the smaller parties. Further, it is not certain that the Tories or the NDP would push strongly for PR in this kind of context: both might imagine that their misfortunes throughout the 1990s have been a temporary setback and that a dramatic rise in their popular support in the future will bring them to the plateau (about 30 percent of the vote or higher, familiar terrain to the Tories) where any party can expect to gain from the plurality system.

The referendum route

The circumstances that led to the introduction of MMP (mixed member proportional system) in New Zealand, and of a mixed system in Italy, have led many supporters of PR to envisage a different scenario, whereby voters would force PR on a reluctant but thoroughly discredited political class. In New Zealand, two referendums held in the early 1990s, the second binding on Parliament, established that the public preferred MMP to the existing plurality system and other alternatives, and despite its own misgivings the government of the day had no choice but to comply. In Italy, voters took advantage of the popular initiative — and of the discrediting of politicians following the tangentopoli scandal — to trigger a referendum on the elimination of the PR element in the Senate’s electoral system, thus forcing parliamentarians to review the electoral systems for both houses.

The New Zealand scenario, which is also thought to be plausible in Britain, is unlikely in Canada because of our historical wariness toward referendums. Moreover, Canada’s political class is reluctant to provide supporters of PR with a golden opportunity to argue their case with the public and possibly win. It does not appear that a referendum on the issue can be organized without the concurrence of Members of Parliament. The existing Referendum Act provides only for referendums on questions “relating to the Constitution of Canada” and even if this wording was stretched so as to include a (non-binding) referendum on PR, the approval of a majority of members of both Houses must be obtained for such a question to be asked.

The Italian scenario is even more unlikely because there is no federal provision allowing citizens to force the holding of a referendum by way of petition. If such a provision existed, as in British Columbia, the introduction of PR would become a possibility. Two crucial variables would condition the effectiveness of the procedure: the number of electors required for a petition to be successful, and whether Parliament would be bound by the result of the referendum (as in Switzerland for constitutional amendments) or would be left free to enact or reject the measure approved by the people (as is the case in British Columbia). A

Charter challenge

A final scenario is the imposition of PR by the courts, in response to a Charter challenge. At first sight, the Charter of Rights and Freedoms seems silent on the issue; the only mention is in section 42(1)(a) of the Constitution Act 1982, which calls for “the principle of proportionate representation of provinces in the House of Commons prescribed by the Constitution of Canada.” This, however, refers exclusively to the redistribution of seats between provinces, not to the representation of parties in the House.

Yet the “right to vote” guaranteed by section 3 of the Charter was given an unexpectedly broad meaning by a British Columbia court in 1989 in the Dixon Case. The court held that it encompassed a “right to equality of voting power” that would be violated by blatant disparities in the population of electoral districts, an approach upheld (though qualified) by the Supreme Court of Canada. The argument might be made, then, that if the right to vote is breached by serious malapportionment of electoral districts which dilutes the voting power of ridings with above-average populations, it is also breached by an electoral system that provides electors who vote for smaller parties with little or no representation.

Nobody can predict how this kind of argument, presented in a more elaborate way, would fare in court. Yet this avenue for change exists, it has recently been explored in greater detail in a law journal,59 and could conceivably be successful in the future.

Conclusions

There is no perfect electoral system, as evidenced by the continued use of both PR and of majority or plurality systems in established democracies, as well as by the spread throughout the 1990s of mixed systems that try to secure the best of both worlds. It is significant, however, that unlike the 1950s, PR systems rarely come under fire nowadays, whereas plurality systems and the Westminster model are frequently challenged, sometimes successfully, in the established democracies where they are used.

On balance, in this country and at this time, the benefits of PR outweigh the disadvantages. Party caucuses would become more balanced regionally, no region would appear to have an overwhelming say within the government party, and the formation of regionally representative cabinets would be facilitated. The chief downside, that large electoral districts would be less suited to constituency work as Canadian MPs have traditionally practiced it, could be offset to a large extent by opting for a German-style mixed system, with 50 percent or 60 percent of members elected in single-member districts.

Our cabinet system would be profoundly transformed by PR. Party elites, rather than the straight will of a plurality of the electorate, would select the government. Single-party majority governments would largely disappear and coalitions would become the standard government formula, though the formation of single-party minority governments would not be ruled out. Cabinets would be less durable, and prime ministers would lose some of the dominance they now enjoy.

Contrary to a widely held view, there is no evidence that PR would necessarily lead to a poorer economic performance or to bad governance more generally. Governments would devote more time to cabinet discussions in order to reach agreement between coalition partners, but the decisions so arrived at might prove to be wiser than some decisions taken impulsively by a prime minister after minimal discussion in cabinet.

For now, the prospects for the introduction of PR in Canada look bleak. While support for some form of PR is strong among academics, there is less consensus among them than in the early 1980s about the best alternative. The public has some sympathy for PR, but the issue does not seem to be an overriding concern. Political circumstances make the implementation of PR by the present government unlikely, while its introduction by the people through a referendum seems improbable. The two most plausible scenarios involve PR being imposed on the leading party, either by a smaller party in a minority Parliament or by judicial decision.

This paper was written before the November 27, 2000 election, which did not substantially alter the pattern of regional support for the parties. As has been the case since the beginning of the 1990s, the performance of the Liberal party throughout the country (172 seats out of 301) again rests on the massive support of Ontario (100 seats out of 103) and masks an even larger deficit in the West (only 14 seats out of 88, one less than in 1997, and 13 less than in 1993). Regional polarization was reduced due to the Liberal advance in the Atlantic provinces and in Quebec at the expense of the Conservatives and New Democrats. However, the Liberals lost ground in the popular vote in each of the four Western provinces. The Canadian Alliance, despite a notable increase in its share of the vote (from 19 percent to 25 percent) did not achieve a breakthrough outside the West: all but two of its MPs are from this region. The Bloc Québécois kept the majority of seats in Quebec by a hair’s breadth, but trailed the Liberals by 4.3 percent (about 151,000 votes) in the popular vote. The complex patchwork of the 1997 elections has been replaced in 2000 with the image of a clear gap between east and west.

Such a result is unlikely to re-ignite the debate on electoral reform at this point. Nevertheless, the fall in support for the Conservatives (from 18.8 percent to 12.2 percent) and for the New Democrats (11.0 percent to 8.5 percent), as well as the inability of the Canadian Alliance to break through in Ontario, reinforces the impression that there exists no serious alternative to the Liberal party at this stage. With at least two seats in each province, this party should, without too much difficulty, be able to construct a cabinet with minimal representation from each region, even if 58 percent of its caucus comes from Ontario and only eight percent from the West. Meanwhile, the turnout rate (at slightly above 60 percent) has fallen to its lowest level since the institution of universal suffrage in this country.

  1. Major contributions in this vein are Alan C. Cairns, “The Electoral System and the Party System in Canada, 1921- 1965,” Canadian Journal of Political Science, Vol. 1, no. 1 (March 1968), pp. 55-80; W.P. Irvine, Does Canada Need a New Electoral System? (Kingston: Queen’s Studies on the Future of the Canadian Communities, Monograph no. 1, 1979); John C. Courtney, “Reflections on Reforming the Canadian Electoral System,” Canadian Public Administration, Vol. 23, no. 3 (Fall 1980), p. 427-57.
  2. See for example Donley Studlar, “Will Canada Seriously Consider Electoral System Reform? Women and Aboriginals Should,” in Henry Milner (ed.), Making Every Vote Count: Reassessing Canada’s Electoral System (Peterborough: Broadview Press, 1999), pp. 123-32.
  3. Ralph Allen, Ordeal by Fire: Canada, 1910-1945 (Garden City, NY: Doubleday and Co., 1961), p. 235.
  4. This finding is included in Paul Howe and David Northrup, “Strengthening Canadian Democracy: The Views of Canadians,” Policy Matters, Vol. 1, no. 5 (July 2000). My thanks to Paul Howe for providing a copy of the draft document prior to its publication.
  5. See Mark N. Franklin, “Electoral Participation,” in Lawrence LeDuc, Richard G. Niemi and Pippa Norris (eds.), Comparing Democracies: Elections and Voting in Global Perspective (Thousand Oaks, Ca.: Sage, 1996), p. 226.
  6. Neil Nevitte, André Blais, Elisabeth Gidengil and Richard Nadeau, Unsteady State: The 1997 Canadian Federal Election (Don Mills: Oxford University Press, 2000), pp. 15-16.
  7. Richard Johnston, “Canadian Elections at the Millenium,” Choices, Vol. 6, no. 6 (September 2000).
  8. See Louis Massicotte, “Parliament in the 1990s,” in J. Bickerton and A.G. Gagnon (eds.), Canadian Politics, 3rd ed. (Peterborough: Broadview Press, 1999), p. 179n.
  9. Richard S. Katz, “Electoral Reform is not as Simple as it Looks,” in Milner (ed.), Making Every Vote Count, p. 108.
  10. We can cite Sweden, where Social Democrats occupied the premiership uninterruptedly from 1932 to 1976, except for a few summer months in 1936; Italy, where Christian Democrats dominated all cabinets from 1945 to the mid- 1990s; and Switzerland, which has been ruled by the same coalition of parties — and the same party balance within the government — since 1959. In Belgium, the Catholic party has been almost continuously in office since 1884, either alone or in alliance with other parties. In Spain, the Socialist party governed from 1982 to 1996 under the same premier.
  11. The limited vote, a plurality formula that tries to ensure representation for the minority party in multi-member districts by granting to each voter a number of votes less than the number of seats in the district, was also tried briefly for electing Toronto representatives to the Legislative Assembly of Ontario in 1886 and 1890.
  12. In the 1997 Canadian federal election, one member was elected with 28.9 percent of the vote. In 1944, in a provincial election held in Quebec, a CCF candidate was elected with 21 percent of the vote. Once in Papua New Guinea, where first-past-the-post applies in a highly fragmented electorate, a candidate was elected in a district with as little as 6.3 percent of the vote.
  13. See André Blais and Louis Massicotte, “Direct Presidential Elections: A World Summary,” Electoral Studies, Vol. 16, no. 4 (December 1997), pp. 441-55.
  14. The number of members elected with 50 percent of the vote or more was 104, including 58 Liberals, 29 Reformers, 11 Bloquistes, 3 Conservatives and 3 New Democrats.
  15. More precisely, alternative voting was in force in Manitoba’s rural districts from 1924 to 1955, in Alberta’s rural districts from 1924 to 1956, and province-wide in British Columbia for the 1952 and 1953 elections. In 1923, a serious attempt was made to introduce alternative voting for Ontario’s rural districts (coupled with a proportional system for urban districts). It failed because a filibuster by the Opposition led the premier to call an early election, which he lost.
  16. Antoine Bilodeau, “L’impact mécanique du vote alternatif au Canada,” Canadian Journal of Political Science, Vol. 32, no. 4 (December 1999), pp. 745-61.
  17. Independent candidate John Nunziata would have been elected in both scenarios.
  18. STV was in force in Manitoba for provincial elections between 1920 and 1955 (in Winnipeg only), and also in Alberta for the election of MLAs from Calgary and Edmonton from 1926 to 1956. Since either the plurality rule or alternative voting were in force at the same time in rural single-member districts, both provinces, in fact, used a mixed system consisting of two distinct electoral systems in different areas.
  19. For a more complete description of the mechanics of the various electoral systems, see André Blais and Louis Massicotte, “Electoral Systems,” in LeDuc, Niemi and Norris (eds.), Comparing Democracies: Elections and Voting in Global Perspective, pp. 49-81.
  20. Louis Massicotte and André Blais, “Mixed Electoral Systems: A Conceptual and Empirical Survey,” Electoral Studies, Vol. 18, no. 3 (September 1999), pp. 341-66.
  21. A survey of the outcomes produced by the various types of mixed electoral systems, using Mackie and Rose’s index of proportionality (from 0 to 100, a higher score meaning a more proportional result) found an average index of 80.5 for superposition systems, much less than the index for plurality systems (87) or majority systems (84). In contrast, the average index for corrective systems was 91.8, quite close to the index for PR systems (94). See Louis Massicotte and André Blais, “Mixed Electoral Systems: A Conceptual and Empirical Survey,” paper presented at the XVII World Congress of the International Political Science Association, Seoul, 1997, pp. 10-11.
  22. André Blais and Louis Massicotte, “Electoral Formulas: A Macroscopic Perspective,” European Journal of Political Research, Vol. 32, no. 1 (August 1997), pp. 107-29. Plurality systems tend to be found in former British colonies, and on the North American, Asian and African continents. Among established democracies, PR is more widespead than the first-past-the-post system, though the latter tends to prevail in larger countries.
  23. A sample includes: David Elton and Roger Gibbins, Electoral Reform: The Need is Pressing, the Time is Now (Calgary: Canada West Foundation, 1980); Ronald G. Landes, “Alternative Electoral Systems for Canada,” paper presented at the 1980 meeting of the Canadian Political Science Association, Montreal, June 2, 1980; Irvine, Does Canada Need a New Electoral System?, pp. 52-58; W.P. Irvine, “Power requires Representation,” Policy Options, Vol. 1, no. 4 (Dec. 1980/Jan. 1981), p. 22; W.M. Dobell, “A Limited Corrective to Plurality Voting,” Canadian Public Policy, Vol. 7, no. 1 (Winter 1981), pp. 75-81; R. Kent Weaver, “Improving Representation in the Canadian House of Commons,” Canadian Journal of Political Science, Vol. 30, no. 3 (September 1997), pp. 473-512.
  24. Full-fledged PR refers here to list systems, but the effects of STV would be quite similar. The chief difference is that in the latter, voters have more freedom to choose from among the various slates of candidates sponsored by each party, though in practice few do so.
  25. For example, New Zealanders were promised the best of all worlds in the early 1990s when the introduction of the mixed member proportional system (MMP) was under debate, and this, combined with almost universal scorn for the existing political class, contributed to the victory of MMP in two referendums. A few years later, the actual operation of MMP had proved so disappointing, even to some of its advocates, that public opinion turned sharply against it, and New Zealand began to be cited as an example of the drawbacks of PR, to the point where people forgot about the real improvements it made in the area of representation. The successful operation of MMP in the 1999 election, however, might swing public opinion back in favour of it again.
  26. In 1993, the Reform Party and the Bloc Québécois, with respectively, 18.7 percent and 13.5 percent of the vote, secured 52 and 54 seats, while the Progressive Conservatives, with 16.0 percent, got only 2. The Bloc reached Official Opposition status while ranking fourth in terms of popular vote. It should not be overlooked that the Bloc fielded only 75 candidates and Reform 207, while the Conservatives had candidates in all 295 ridings.
  27. See Wilma Rule, “Women’s Underrepresentation and Electoral Systems,” Political Science and Politics: PS, Vol. 27, no. 4 (December 1994), pp. 689-93.
  28. Lord Jenkins of Hillhead (Chair), The Report of the Independent Commission on the Voting System (London: HMSO, 1999), para. 39.
  29. Since 1993, Ontario has been a one-party province in terms of representation in the House of Commons, with the Liberals sweeping all but one of Ontario’s 99 seats (1993) and all but two of 103 seats (1997), though the Liberal share of the Ontario vote did not exceed 53 percent in either case.
  30. In 1997, 40 percent of Reform’s votes came from non-Western provinces (35 percent from Ontario alone), but this did not translate into any seats.
  31. Belgium initially had a two-ballot majority system in multi-member districts. In the 1894 election, before PR was introduced, the Catholic party won all the seats in Brussels and 71 of the 72 Flemish seats, but only 14 of the 62 seats in Wallonia. See Xavier Mabille, Histoire politique de la Belgique (Bruxelles: CRISP, 1986), p. 194. This result illustrates how little a two-ballot system can do to mitigate regional polarization.
  32. Canada, Task Force on Canadian Unity (Pépin-Robarts): A Future Together: Observations and Recommendations (Ottawa: Ministry of Supply and Services, 1979), p. 105.
  33. See Massicotte and Blais, “Mixed Electoral Systems: A Conceptual and Empirical Survey,” pp. 341-66.
  34. Leigh J. Ward, “Second-class MPs’? New Zealand’s Adaptation to Mixed-Member Parliamentary Representation,” Political Science, Vol. 49, no. 2 (January 1998), pp. 125-45. This source, the only one we know about that mentions this kind of problem with mixed systems, concludes that “the problem is less one of reality than of inaccurate perception” (p. 143).
  35. See David W. Brady et al., “The Decline of Party in the U.S. House of Representatives, 1887-1968,” Legislative Studies Quarterly, Vol. 4, no. 3 (August 1979), pp. 381-407.
  36. Louis Massicotte, “The Rise of Party Cohesion in the Canadian House of Commons 1867-1945: A Descriptive and Comparative Summary,” paper presented at the Third Workshop of Parliamentary Scholars and Parliamentarians, Wroxton College, Oxforshire, UK, August 1998; Joseph Wearing, “Tweaking the Whips: Modified Rebelliousness in the Canadian House of Commons,” paper presented at the Third Workshop of Parliamentary Scholars and Parliamentarians, Wroxton College, Oxforshire, UK, August 1998.
  37. Jean-Claude Colliard, Les régimes parlementaires contemporains (Paris: Presses de la Fondation nationale des sciences politiques, 1978), pp. 210-16.
  38. Louis Massicotte, “Bridled Workhorses: Party Discipline in Committees of the Canadian House of Commons,” paper presented at the annual meeting of the Canadian Political Science Association, St. John’s (Nfld.), June 1997.
  39. The elections of 1958, 1968 and 1974 provided an incumbent minority government with a majority of seats.
  40. Michael Laver and Norman Schofield, Multiparty Government: The Politics of Coalition in Europe (Oxford: Oxford University Press, 1990), p. 70.
  41. See for example Jonathan Boston, Governing under Proportional Representation: Lessons from Europe (Wellington: Institute of Policy Studies, University of Victoria, 1998).
  42. Laver and Schofield, Multiparty Government, p. 152. The median duration of minimum winning coalitions (35 months) was higher than the mean.
  43. Laver and Schofield, Multiparty Government, pp.148-49.
  44. For example, Tage Erlander served as Prime Minister of Sweden without interruption from 1946 to 1969, a period of time (23 years) that exceeds that served by Mackenzie King, Canada’s longest-serving Prime Minister. In recent years, Felipe Gonzalez was Premier of Spain for 14 years, while Helmut Kohl led Germany for an uninterrupted 16 years.
  45. From 1945 to 1995, there were 42 cabinets in Israel (starting from 1948), 46 in Finland and 55 in Italy. The French Fourth Republic saw 28 cabinets in its 13 years of existence. See Jaap Woldendorp, Hans Keman and Ian Budge, “Party Government in 20 Democracies: An Update (1990- 1995),” European Journal of Political Research, Vol. 33, no. 1 (January 1998), pp. 125-64.
  46. See Donald J. Savoie, Governing from the Centre: The Concentration of Political Power in Canada (Toronto: University of Toronto Press, 1999).
  47. One was attempted against John Diefenbaker in 1963, but failed.
  48. Australian political history offers at least two examples of this. After the 1922 election, the Country party, whose support had become a precondition for the continuation in office of the Nationalist party government, insisted not only on an almost equal number of cabinet ministers, but also blackballed incumbent Prime Minister Hughes, thus obliging the Nationalist party to select a new leader as Prime Minister. Similarly, in 1967-68, following the death of Liberal Prime Minister Holt, Country party Leader John McEwen vetoed the appointment of William McMahon as Prime Minister, on the grounds that McEwen did not trust him personally.
  49. Arend Lijphart, “Democracies: Forms, Performance, and Constitutional Engineering,” European Journal of Political Research, Vol. 25, no.1 (January 1994), pp. 1-17; and Patterns of Democracy: Government Forms and Performance in Thirty-Six Countries (New Haven and London: Yale University Press, 1999).
  50. Lijphart, “Democracies: Forms, Performance and Constitutional Engineering,” p. 12.
  51. Kaare Strom, Minority Government and Majority Rule (Cambridge: Cambridge University Press, 1990), p. 238.
  52. Strom, Minority Government and Majority Rule, p. 238.
  53. See Louis Massicotte, “Federal Countries, Elections in,” in R. Rose (ed.), International Encyclopedia of Elections (Washington, DC: Congressional Quarterly, 2000), p. 101-04.
  54. Louis Massicotte and André Bernard, Le scrutin au Québec: Un miroir déformant (Montréal: Hurtubise HMH, 1985).
  55. Herman Bakvis and Laura G. Macpherson, “Quebec Block Voting and the Canadian Electoral System,” Canadian Journal of Political Science, Vol. 28, no. 4 (December 1995), pp. 659-92.
  56. André Blais and Stéphane Dion, “Electoral Systems and the Consolidation of New Democracies,” in Diane Ethier (ed.), Democratic Transition and Consolidation in Southern Europe, Latin America and Asia (London: Macmillan Press, 1990).
  57. Thomas Ertman, “Democracy and Dictatorship in Interwar Western Europe Revisited,” World Politics, Vol. 50, no. 3 (April 1998), pp. 475-505.
  58. This section is based on the following documents, as they appeared in July 2000 on the Internet sites of political parties. Canadian Alliance (Reform Party): Policy Declaration, para. 74 (Site http://www.reform.ca); Progressive Conservative Party: Policy Document adopted by the national convention in May 2000 (Site http://www.pcparty.ca); New Democratic Party: The Social Democratic Forum on Canada’s Future, a document adopted in August 1999 by the party’s Federal Convention (Site http://www.ndp.ca); Bloc Québécois: Orientations adopted by the party at its convention held in January 2000 (Site http://blocquebecois.org); Green Party of Canada: Policy adopted by the 1998 convention (Site http://www.green.ca).
  59. Trevor Knight, “Unconstitutional Democracy? A Charter Challenge to Canada’s Electoral System,” University of Toronto Faculty of Law Review, Vol. 57, no. 1 (Winter 1999), pp. 1-42.