Case for proportional representation in Canada.

Henry Milner is a political scientist living in Montreal and a co-editor of Inroads. He spent the fall of 1996 in New Zealand studying the implementation and effects of proportional representation.

A bit of history

The principles of electoral democracy had been accepted by 1867, when three of the remaining British colonies in North America federated to form the Dominion of Canada. This was the same year Britain extended its suffrage to 10 percent of the electorate. While Canada's founding fathers accepted a form of American-invented federalism, they, in contrast with their Australian counterparts two generations later, took for granted the electoral system inherited from Britain, failing to ask if it suited a federal country dispersed over farflung regions. As Tom Flanagan points out, provincial elections in the West were conducted under different systems of election earlier in this century, but today, not only are the 301 members of Parliament elected through FPTP, but so is every member of the 10 provincial legislatures and two territories. Indeed, the federal electoral system moved even more closely to a pure FPTP model as the few two-member districts that once existed were eliminated. (1)

The appropriateness of FTPT for Canada has been largely taken for granted, in part because Canadian familiarity with electoral experiences outside its borders generally extends only to the U.S. and U.K. Yet this does not fully explain how a country so concerned with constitutional reform has not been open to altering its electoral institutions--especially, as we shall see, given the anomalies they have produced.

This is not to say that reform to a more proportional system has never been proposed, only that it has not made it to the political agenda. The Task Force on Canadian Unity (Pepin-Robarts Commission) in its 1979 report included a recommendation for just over 20 percent of the seats to be accorded to the parties proportional to their electoral support, to enable them to gain representation from provinces denied to them by FPTP. A slightly different proposal was submitted by the NDP, the party most under-represented under FPTP. But when the Trudeau government unceremoniously rejected the Pepin-Robarts report, electoral reform of the House of Commons was also shelved.

That the issue was off the political agenda was underlined when Pierre Lortie, Chairman of the Royal Commission on Electoral Reform and Party Financing set up by the Mulroney government in 1990, made it clear that changing the electoral system as such was outside the Commission's mandate. Discussion of electoral reform of federal legislative institutions was now limited to the proposal to have an elected Senate. But Senate reform--and with it, electoral reform--died with the rejection of the Charlottetown constitutional amendment proposal in the 1992 referendum. Ironically, the distorting effects of the FPTP electoral system on representation in the House of Commons have probably never been greater than in the two federal elections that followed.

How FPTP can distort electoral outcomes

The results of Canada's last two federal elections are becoming political science textbook cases of the distortions under FPTP. In 1993, the voters repudiated the ruling Progressive Conservatives, but the electoral system almost decimated Canada's oldest party. Rather than electing the 46 members a proportional system would have given them, the Tories managed to elect only two. In contrast, the two regionally based parties, the Bloc Quebecois and Reform, with 13.5 and 19 percent of the popular vote respectively, elected 54 and 52 members.

In 1997, of the 301 seats in Parliament, the Liberals won 155, Reform 60, the Bloc Quebecois 44, the NDP 21 and the Tories 20. The regionalization of the outcome was even more pronounced in 1997. Two thirds of the Liberals' seats came from Ontario, where their 48.5 percent of the vote gave them 101 of 103 seats. Reform dominated the western provinces, the Bloc Quebecois Quebec, and the Conservatives and NDP did best in the Atlantic provinces--"quartering Canada" as The Economist put it, and producing what Canadian pundits called a "Rainbow Parliament." As most editorialists noted at the time, the Liberal government owed its--bare--majority to a sweep of Ontario and English Quebec, its red bastion encircled by Reformist green to the west and Bloquiste bleu to the east. But, as the more perceptive of the editorialists also noted, the "quartering" was inflicted more by FPTP than by Canadian voters. If the election had been fought under proportional representation (PR), the result would have been much different.

Had the seats been distributed according to the parties' popular national support, the Conservatives would have placed third with 58 seats, just behind Reform's 59, with the NDP up and the Bloc Quebecois down to 33 each. The Liberals would have been left with 118 seats. Liberals, Conservatives and NDPers would have won seats in all provinces or regions, and Reformers in all but Quebec. If, as would be the case, re-allocation were made based on the provincial vote, and depending on the exact allocation method used, the Greens--as the table below based, on the St-Lague method, shows--might have elected a member in both Ontario and British Columbia.

But the contrast between a proportional Parliament and Canada's "rainbow" Parliament would have been even greater than these overall numbers suggest, for the campaign itself would have been very different in an MMP environment. The parties would have had no incentive to concentrate their efforts and resources in the regions where they were doing well, since, in contrast with our present system, under PR every vote counts equally toward electing an MP. The Conservatives would have put far more effort into the west. The NDP and Reform would have worked harder for support in Quebec, and would likely have been rewarded with seats for their efforts. The Bloc might even have been tempted to run candidates outside Quebec, since any votes garnered could only help it win seats.

         BC  Alta  Sask  Man  Ont  Que  NB  NS  PEI  Nfld  Terr  Total


Liberal  10   6     4     5   51   27   3   3    2    3     1     115

PC       2    4     1     3   19   17   4   4    1    3     1     59

Reform   15   14    5     3   20        1   1                     59

NDP      6    2     4     3   11    2   2   3    1    1     1     36

BQ                                 29                             29

Green    1                     1                                   2

Other                          1        1                          2

Total    34   26    14   14   103  75   10  11   4    7     3     301

Source: a calculation by Julian West distributed on the Electoral Reform

Website

Voters in these regions would have been far more receptive to such appeals since they knew they would not be wasting their votes. Indeed, there can be no doubt that the dismal turnout of only two-thirds of registered voters is linked to the fact that, in most ridings, only one or two of the parties were real contenders, with supporters of the others effectively disenfranchised. (2) Moreover, where you seek votes affects what you say. Once parties concentrate less on regional strongholds and more on the country as a whole, they have every incentive to moderate the divisive elements of their platform, and emphasize the unifying ones. Clearly Reform--and maybe even the Bloc--would have sung a different, more harmonious, tune under PR.

The MMP alternative

We have compared existing electoral outcomes to those that would have resulted from an election under PR. It is now time to be specific about the PR system best suited for Canada. Without repeating the arguments, I propose we endorse what was the clear preference of the New Zealand Royal Commission, and overwhelmingly ratified by the population in the 1992 referendum:

The German Electoral System

When a general election is held, half the members of the German Parliament--the Bundestag--are nominated in local party meetings and elected from 248 single-member constituencies (Wahlreise), in which a simple plurality suffices for election. In what is often referred to as the "second vote," the other half of the Bundestag members are nominated in state-wide conventions and are elected from a party list in each Land (state). These members of the German Bundestag are elected from multi-member party lists, an electoral structure commonly associated with European parliamentary systems. The list of candidates cannot be altered; no preference voting exists. Candidates may run for office in either or both electoral systems.

Seats to the Bundestag are allocated in the following manner: first, the percentage of votes a party receives in the second vote determines the overall number of seats to which that party is entitled--unless it receives less than 5 percent, in which case it receives no seats whatsoever. In other words, a party that receives 25 percent of all the second votes will receive 25 percent of parliamentary representation. Somewhat misnamed, the second vote is the most important to the party. The party list is critical to the competing political parties, because a dropin votes here directly translates into a loss of parliamentary seats. Second, seats are given to all individuals who won a plurality in the Wahlkreis. Third, returning to the second vote, individual candidates are selected from the party list, in the order listed by the parties, until the overall percentage or proportion of seats to which that party is entitled, based on the second vote, is allocated. Normally, there will also be 248 seats awarded to candidates from party lists, although occasionally additional seats, referred to as an "overhanging mandate," are allocated to maintain proportionality...

Source: Thomas D. Lancaster and W. David Patterson, "Comparative Pork Barrel Politics: Perceptions from the West German Bundestag," in Comparative Political Studies, Vol. 22, No. 4, January 1990, pp. 458-477. the German-style system known as "mixed-member proportional." MMP gives each party that achieves at least five percent of the popular vote an overall number of seats proportional to its share of the popular vote. However, since half those seats come from single-member districts, each citizen still has his or her own MP.

The German electoral system is described in the insert at left. Apart from the five MPs elected in separate Maori districts (an idea that could be usefully explored in Canada), the electoral system in unitary New Zealand differs from that in federal Germany only in that the parties provide single national lists for the party vote and not provincial (Lander) lists. Given Canada's federal system and far-flung territory, we would be wise to keep this aspect of the German system if we were to adopt MMP in Canada. Despite this, I refer mainly to New Zealand in this article, since New Zealand is, like Canada, a country with inherited "Westminster" institutions and traditions.

I choose MMP over the various list-based systems used in much of Europe mainly because Canadian voters--like those in New Zealand--would be unlikely to accept an electoral system that deprived them of having a single MP to represent them. Among PR systems, only MMP achieves this. I suspect also that Canadians would insist on a high threshold, like the 5 percent in Germany and New Zealand, which would limit the number of parties winning seats to, typically, four to six. I do not discuss the option whereby voters exercise some preference among their preferred party's listed candidates, (as is done in a number of PR countries). It could be incorporated into MMP should Canadians prefer.

Forming a government under MMP

The moderating tendency that PR has on party campaigning affects government formation as well. Rather than being a force for disunity, as Jean Chretien warned, a Liberal minority could well have been a stronger force for unity than the present majority, so rooted in central Canada. With just under 40 percent of seats spread over the entire country, the Liberals would have had the choice of ruling as a minority or forming a coalition with another party.

Opponents of PR warn us of the dangers of such outcomes. Yet a minority Liberal government would have had little fear of being defeated in Parliament since the other four parties have so little in common. (A situation in which opposition disunity allows for long-term, stable minority government is not all that uncommon these days--just look at Norway--and try to imagine the wording of a vote of non-confidence that would have been supported by the Bloc, Reform, Tories and NDP.) Of course, a minority government would have had to work harder, having to turn to its left or its right, to centralizers or decentralists, for support in Parliament--depending on the legislation in question. The probable outcome would not only have been better legislation, but legislation with a degree of popular legitimacy that numerical majority government cannot achieve.

We often look back upon the postwar decades as a period of good government. Yet of the nine federal elections that took place between 1957 and 1979, six resulted in minority governments. Faced with the prospect of a possible minority government, the Canadian Press, on May 30, 1997, tracked down what the memoirs of Pierre Trudeau (who was a minister in one minority government, led another, and headed the Opposition in a third) had to say on the subject: "They were exciting times, akin to canoeing through seething rapids. A leader learned how to live dangerously, how to savor the pleasures of running risks and overcoming perils... If you can't do that when you are in a minority government, you shouldn't be in politics."

But this was under FPTP, when minority governments were regarded as exceptional. Under MMP, the absence of a majority government would be the norm. Hence, as in most PR countries, we might well have found ourselves with a coalition government. One possibility would have been a Liberal-Conservative coalition, which would have far greater legitimacy in dealing with the national unity question than does the present government--a minority government in everything but the number of seats--and one in which the national unity dossier could have been in the hands of Jean Charest.

I stress the effect on national unity since the issue dominates Canadian federal politics. But the main effect of PR is to temper ideological swings. Consider the last two Ontario elections, each of which resulted in governments ideologically more extreme than the majority of Ontario voters.

The second produced Mike Harris's "common sense revolution," a series of radical policies reminiscent of those of New Zealand's National (Conservative) Party from 1990 to 1993. But New Zealanders, in their wisdom, understood that electoral institutions needed to be changed if they wished to avoid narrow, ideology-driven agendas imposed by a single-party majority. In 1993, New Zealanders narrowly re-elected National but endorsed MMP in a referendum. National took little time to open itself to more centrist policies, policies implemented under the present centre-right coalition government elected under MMP in 1996. While the transition has not been painless, New Zealand finds itself with small parties of the right, left and centre, and large parties of the centre-right and centre-left, guaranteeing--as much as such things can be guaranteed--that the government will reflect the considered will of the majority of its citizens.

Political institutions shape our politics and policies

We now have years of experience in the large majority of democratic countries that use PR, and the political science literature, notably the work of Arend Lijphart and his colleagues, is quite definitive. PR systems elicit higher voter turnout and representation of women and minorities, and are, on balance, more effective in providing government performance that is both efficient and linked to commitments made in electoral campaigns. (3)

This is no accident. By assuring that the number of seats that parties are accorded will reflect the parties' popular support, PR frames incentives and disincentives for political actors resulting in a reduction of the cost of political information. With disparities in their support not exacerbated by the electoral system, political parties know that they will have to co-operate to govern, and that part of that co-operation consists of undistorted transmission of information among political actors. Where compromise and coalition is a visible, built-in feature of the political process, opponents can collaborate even when they disagree.

Moreover, PR works against distortion in the flow of information from top to bottom by reducing the cost to political leaders of making the electorate aware of alternative positions on salient policy options, and of how they, as opposed to their opponents, view their likely effect if implemented. Under FPTP, in contrast, the governing party is expected to implement its program as if supported by a majority of the population, rather than seek and build broad-based support for needed, but controversial, reforms. It knows that the other parties have nothing to gain by co-operating, that their political interest lies in denouncing, distorting or exaggerating the likely effects of unpopular policies, even ones they know to be necessary. And, in a world, of global markets, high debt and environmental and demographic crises, such policy choices will more and more confront a misinformed (4) and increasingly alienated electorate--making FPTP increasingly a luxury we cannot afford.

If we are concerned with preserving the relative equality of the welfare state against these forces, we should be thinking about the institutions most conducive to such

Income Distribution in Rich Democracies

Country        Gini    Year    Gini    Year  PR
              (80s)           (90s)


Finland         .207    87      .227    91    +
Sweden          .220    87      .229    92    +
Austria         .227    87                    +
Norway          .234    86      .230    91    +
Denmark                         .239    92    +
Belgium         .235    88      .230    91    +
Italy           .310    86      .255    91    +
Luxembourg      .238    85      .268    91    +
Germany         .250    84                    +
Netherlands     .268    87      .268    91    +
Canada          .289    87      .285    91    -
Australia       .295    85      .309    90   -/+
France          .296    84                   -/+
U.K.            .304    86      .335    91    -
Switzerland     .323    82                    +
U.S.            .341    86      .350    91    -
N.Z.         .26(.23)   86   .33(.30)   92    -

Source: Luxembourg Income Survey

outcomes. In my books and articles on social democratic welfare states, I have explained why the social compromises fostered by PR are more likely to arrive at such outcomes. Here, I will let the numbers speak for themselves.

The table at left sets out the most recent Gini coefficient ratings for Western countries. The Gini coefficient measures the level of equality or inequality in the distribution of revenues among the poorest to richest deciles of the population. At one theoretical extreme, a Gini of 0.0 indicates that the 10 deciles of households in a given society each have the same total disposable income, while, at the other extreme, a Gini of 1.0 indicates that the richest decile has all the disposable income. In recent years, the Luxembourg Income Survey (LIS) has standardized the methodology so that the LIS-assembled Gini data now available for most industrial democracies gives us a solid, long-term basis for comparison. While there are partial exceptions--notably Switzerland--the rule is that PR countries are the most egalitarian and FPTP countries (the U.S., Britain, New Zealand) the least; countries with systems in between, like France and Australia, are in the middle. (New Zealand's figure are not from LIS, but from sources approximating the LIS methodology.

Time for a Canadian debate

Following the last two federal elections, electoral reform towards a more proportional system was proposed by a number of columnists and editorialists, and raised by leaders of the Progressive Conservative Party, but only wistfully. In November 1997, a private member's bill was submitted by Lorne Nystrom of the NDP proposing Parliament endorse PR and appoint an all-party committee to conduct public consultations and report back with a concrete proposal, which would then be put to Canadians in a national referendum. But like other private members' bills, this one--even if it wins the lottery and makes it to the floor of the House--will die on the order paper. Despite the fact that the 1997 election result made apparent to NDP and Conservative leaders the disadvantages of FPTP, politicians still view electoral reform as a non-starter in which they are unwilling to invest precious political capital.

While this is understandable it is also regrettable. Where the FTPT system has produced majority governments, the tendency of the system to polarize rather than promote compromise has not necessarily served Canada well. As a thought experiment, one might imagine the outcome if the one serious recent effort at electoral reform had succeeded.

In the early 1980s, a Quebec investigatory commission advocated adoption of a regional-list system of PR. The recommendation was endorsed by the Levesque Cabinet, but, due to lack of support, was never presented to the legislature. (5) Had it been adopted, the balance of power today would be held by parties representing the 30 percent of Quebecers who insist on a new relationship with Canada but prefer a compromise short of sovereignty. The National Assembly would be made up of at least four parties: hard-line independentistes like Jacques Parizeau; moderate sovereignists, probably led by Lucien Bouchard; nationalist federalists like Claude Ryan; and Trudeau federalists led, perhaps, by Guy Bertrand or an anglophone from the Equality Party. If PR were adopted before the next Quebec election, the Action Democratique, whose leader Mario Dumont has called for a 10-year moratorium on referenda in Quebec, would likely--Jean Charest or no Jean Charest--win the balance of power. Of course, despite the fact that electoral reform is still in the PQ program, it will remain off the political agenda at least until after the next referendum.

Prospects for reform

Currently, the main electoral reform efforts centre around recall, a measure that, if adopted widely, could undermine informed democratic party politics, just what moving to MMP would enhance. Nevertheless, the sentiment behind the recent effort in B.C. to use recall against two provincial legislators and thus undermine the paper-thin majority of the NDP government is quite understandable--after all, the NDP won with many thousands of votes fewer than the opposition Liberals. But recall treats only the symptoms, with grave side effects.

Considering major electoral reform is one thing; implementing it another. As Kent Weaver reminds us, it is never easy to change an electoral system, since politicians have a vested interest in maintaining the system that elected them. Apart from the fact that few Canadians' horizons stretch beyond the U.S. and Britain, the explanation for Canada's failure to even discuss change probably lies in a general sense of institutional vulnerability as far as the federal distribution of power is concerned--an unarticulated fear that tampering with electoral institutions would only exacerbate the situation. Nevertheless, recent election results did provoke some positive editorial comment on electoral reform. Outspoken Southam columnist Andrew Coyne reacted to the election results with a call for replacing an electoral system that encourages "the exploitation of regional grievances--grievances to which the same electoral system, by shutting out one region or another from the councils of power, lend real substance [and that displaces] healthy ideological debate in favour of corrosive regional splits." More representative, however, was Jeffrey Simpson of The Globe and Mail who, just after the election, wrote two columns which dismissed PR as "a delicate flower that will wilt again," and which, in the unlikely case of its implementation, would "make regionalism worse." The fact that he is wrong on this last count, as our simulation based on the 1997 federal election shows, does not weaken his influence or the representativeness of his views.

While I would settle for the moderate changes advocated by Weaver, Flanagan and others, I would not want to do so without a public discussion in which MMP gets a fair hearing. For only a truly proportional system would constitute a full break from FPTP, with its built-in polarization--us versus them, the "ins" who have all the answers, versus the "outs" who denounce them all--and towards a system capable of finding intelligent compromises based on the informed wishes of the majority.

Is it likely? Clearly not. But since we have slavishly followed the British in our electoral institutions, perhaps we might follow them again. The new 108-member Northern Irish Assembly is to be elected by PR as one means of ensuring the representation of all communities and all views. And members of the new democratic Assemblies for Scotland and Wales will be elected through a moderate form of MMP. Moreover, it may not end there: as Andrew Reynolds tells us, Britain is contemplating changes to elections to the hallowed halls of Westminster itself.

Who would have expected New Zealanders to elect a government under MMP in 1996? Public opinion would not allow politicians to shelve a Royal Commission report that recommended MMP--public opinion informed by experience and by the careful analyses of political scientists like Peter Aimer. Aimer's conclusion as to why New Zealand was ripe for reform applies even more clearly to Canada: that in the context of a rise in voter volatility, a weakening of the electoral alignments between parties and social groups, and the development of more pluralistic social structures, the plurality-based, two-party system is fragmenting into a more multi-party format more compatible with a proportional electoral system.

(1) . This article considers only the effects of electoral systems in mature democracies. But systems like FPTP have even more acute effects in developing countries. They foster clientalism--representatives dependent on the vote of constituents who care only about economic pay offs and not about their wider costs. Consider Thailand's recent problems. As Ammar Siamwella noted in a magisterial public lecture at Queen's University last fall, Thailand's clientalist political leaders could do nothing about the financial crisis caused by technocratic incompetence. Indeed, the problem was beyond their comprehension. Hope, he concluded, lies in the new Constitution's anticipated change to the complexion of the new Parliament: "Alongside 400 members elected from individual constituencies, there will be a further 100 members belonging to party lists to be elected nationally. ... It is likely that each party will have on this list individuals whom it expects to nominate to be ministers [who will be] expected to take a less localistic stand on issues and initiate debates on more national issues."

(2) . This result brings Canada near the bottom among Western democracies in voter turnout, something we seldom notice since we tend to judge ourselves only against the U.S., where turnout is 50 percent or less. Yet, even here, we overestimate our relative performance since the U.S. calculation is based on potential rather than registered voters. According to the latest Handbook of the IDEA of potential Canadian voters, an average of 69.2 percent voted in federal elections between 1945 and 1993. In the 1997 election, 60 percent voted.

(3) . Despite expectations to the contrary, the "manifesto group" found, in its detailed study of party programs and government policies in 10 Western countries, that PR-based parties are at least as good at turning their programmatic commitments into government policy as those in FPTP-based systems (See Richard I. Hofferberg, "Parties, Policies, and Democracy: An Overview" in H.D. Klingemann, R.I. Hofferberg, and I. Budge, eds., Parties, Policies, and Democracy, Boulder: Westview, 1994). It's time to dispense with the notorious Italian red herring on this matter. Italy's problems of attaining stable governments had little to do with PR per se. To the extent that political institutions and not deeper divisions were to blame, these lie in such specific Italian practices as secret ballots in Parliament and preference voting within party lists, as well as effectively no minimum to discourage tiny fringe parties. It is these practices which undermine stable coalition government, rather than the PR system itself. (Another argument I recently heard bears mentioning here, namely that PR is unacceptable since, under Weimar, it brought Hitler to power. Flabbergasted as I was by the assertion, I failed to note in reply that this "fact" did not discourage Israel from adopting PR in 1948.)

(4) . For the link between information and electoral institutions, see Henry Milner, "Electoral Systems, Integrated Institutions, and Turnout In Local and National Elections: Canada In Comparative Perspective," in Canadian Journal of Political Science, (1997) 30,1, pp. 89-106. A useful illustration of the effects of lower levels of political knowledge on electoral choice is provided in a 1998 book by Donald Granberg and Soren Holmberg, The Political System Matters: Social Psychology and Voting Behavior in Sweden and the United States (Cambridge). It is the most politically ignorant American voters that disproportionately switch sides during an election campaign. This is simply not the case in Sweden.

(5) . For a discussion of the aborted electoral reform effort in Quebec in the early 1980s, see Henry Milner, "Obstacles to Electoral Reform in Canada," in the American Review of Canadian Studies, (1994) 24,1, pp. 39-55.