The mixed member/proportional system used in Germany and New Zealand is sometimes oversold, but it might help lessen some of the imbalances currently observed in Canada.

Fundamental electoral reform, as distinguished from mechanical tinkering, is currently attracting more attention than is usually the case. There are several reasons for this:

  • A new party, the Greens, is emerging and making inroads, winning 12 per cent of the vote at the most recent BC election.
  • Many perceive our first-past-the-post system (FPTP) as harmful for the cohesion of the Canadian federation. In recent federal elections, regional variations in party support have been translated into sharp cleavages in parliamentary representation. The ruling party caucus is dominated by its massive Ontario wing, with few members coming from the West. It becomes difficult to find suitable ministers from each province.
  • There has been a succession of dysfunctional election outcomes in recent years. In Prince Edward Island, British Columbia (2001) and Alberta, the opposition has been reduced to numbers that hardly allow it to fulfil its role. In British Columbia (1996) and Saskatchewan, elections have resulted in governments that were not supported by a plurality of the vote. In Quebec, the excessive concentration of federalist and Liberal party support in English-speaking ridings has resulted in a huge and systematic bias in favour of one side. Three times in a row within the past six years (the 1995 referendum, the 1998 provincial election and the 2000 federal election), nationalists won a majority of districts while trailing in the popular vote. Projections made on the basis of the 1998 results suggest that to win a bare majority of seats, the Charest Liberals need a 7.5-point lead over the PQ—some 300,000 votes. A deliberate gerrymander could hardly produce a more effective result.
  • There is increasing concern among opposition parties, and among less committed observers, that the federal Liberals could hold on to power for a very long time simply for lack of a plausible alternative. The Mulroney coalition has fragmented into three parties. Despite strenuous efforts, including a change of name and leader, the Canadian Alliance has so far proved unable to expand beyond its Western bailiwick. The Bloc Québécois is unwilling, even if it were able, to expand beyond its Quebec stronghold, and is stalling, though it seems unlikely to crash. The essentially negative coalition that killed the Charlottetown Accord has so far been unable to present a viable alternative. While this might work to the advantage of the Tories, the latter have difficulty overcoming the stigma attached to the Mulroney years: With regionally balanced, but weak, support in the country, they have been, and are bound to remain, with the NDP, the main victims of the present electoral system. This has allowed the Chrétien Liberals, thanks to the massive support of Ontario, to maintain at each election (and in all opinion polls conducted in between for the past eight years) a healthy lead over each alternative. Hence the opinion that the electoral system should provide an incentive for the “outs” to conclude electoral alliances in order to get in.

These concerns are not necessarily shared by everybody, and do not lead by themselves to a single reform option. So far, two broad avenues have been proposed: Australian-style alternative voting (AV), or some form of proportional representation (PR), either straight PR in huge multi-member districts, or a German-style mixed system. These two alternatives deserve closer examination.

The reform option that would have the most modest consequences is AV. It would be unlikely to produce more proportional outcomes, and would hamper the emergence of smaller parties. It would keep single-member districts, but allow the losers to coalesce wherever the winner has been supported by less than 50 per cent of the vote. On that basis, many dream of a Tory Alliance informal coalition that would rout the Grits. The likelihood of such an outcome is probably overstated, however.

True, AV would facilitate the conclusion of electoral alliances, but those who assume that Tory and Alliance votes would mechanically combine votes and kick out the Grits are bound to be disappointed, for the following two reasons:

  • Both Australian and Canadian experience suggests that AV makes minimal difference compared with the outcomes produced by FPTP. Take the most recent (1998) Australian election. In 50 districts, the leading candidate got a majority of the vote and won, so that second preferences were not even examined. Further, in the vast majority (91 out of 98) of districts where subsequent preferences were counted, the transfers simply confirmed the victory of the candidate leading after the first count. Finally, in those very few districts—seven, to be precise—where preferences overturned the results of the initial count and resulted in the defeat of the leading candidate, the changes tended to cancel each other out, with the net impact (two seats taken by Labour from the Coalition, out of a total of 148) being pathetically small.
    Was this a mere accident? Not at all: the 1998 election fits the standard pattern. Between 1919 and 1996, there were 21 general elections in Australia, involving 3540 seats. In 70.5 per cent of cases, the first count was conclusive and in a further 1.7 per cent, the winning candidate was acclaimed. In the 983 districts where preferences were counted, they most of the time confirmed the victory of the leading candidate and changed the result in a minority of cases. On the whole, only 213 leading candidates (barely six per cent of the total) were defeated by the distribution of second preferences. The same kind of results were recorded in Australian states and in those Canadian provinces where AV existed for a while.
  • Projections suggest further that in Canada the change wrought by AV would not go in the expected direction, and indeed would increase, rather than decrease, the number of Liberal seats. Antoine Bilodeau’s projection, based on the Canadian 1997 Election Study results, gave 57 per cent of the seats to the Liberals rather than 51 per cent. Why? Simply because the work of AV is more complex than many assume. If we set aside the voters who supported marginal parties, the chief reservoirs of second preferences are the NDP and Tory voters, because they voted for the parties that are the most likely to be eliminated throughout the counting process. NDP voters are more likely to support the Liberals than the Alliance. As to Tories, they are a much more complex group than right-wing ideologues assume. Some are closer to the Alliance, but many others, and not only in Quebec, have a national rather than sectional orientation and consider the Liberals the lesser evil. Further, in Quebec, AV is unlikely to work against the Liberals. For all their grievances against Jean Chrétien’s brand of federalism, Tory voters are mostly federalists and they, too, consider the Liberals as a lesser evil—as evidenced by the 2000 election, in which most former Tories switched to the Liberals rather than to the Bloc or the Alliance.

Let us sum up. AV is an attractive proposal if you want to change as little as possible, if you think that all other concerns raised by the supporters of electoral reform—regional balance and the insufficient representation of women and smaller parties—are unimportant, and if you believe that opposition parties should have an opportunity to unite and rout the Grits. But even on the latter point, the proponents of AV are likely in for a nasty surprise.

What, then, about a more radical change like PR? In a paper published earlier this year by the IRPP, I identified two different types of PR and tried to gauge the impact of each.

In this article, I will leave aside the first type, continental European-style “full-fledged PR,” in which the whole membership of the House of Commons would be elected in huge constituencies (regional districts in Ontario, Quebec, British Columbia and Alberta, province-wide districts elsewhere) according to the percentage of the votes cast for party lists. Apart from the fact that few people advocate a break with our electoral customs of such great magnitude, I strongly believe that an arrangement of that nature is impracticable in the Canadian context, at least outside metropolitan areas, simply due to the sheer size of the country: Imagine a “regional” district encompassing the whole of Northern Ontario! Further, I agree with MPs that single member districts are a better framework for constituency casework, which is what most MPs do and many find most rewarding. A few formerly “straight PR” countries have acknowledged this in recent years by switching to mixed systems where half or more of the membership of the legislature comes from single-member districts, in order to provide for closer links with constituents.

This leaves us with a “German-style system,” known in New Zealand as “mixed member proportional” (MMP). In this kind of hybrid system, half (or 60 per cent) of MPs would continue to be elected by plurality in (admittedly enlarged) single-member districts. However, the remaining seats would be distributed province-wide in a corrective way (“topping up” in Lord Jenkins’ jargon) so that in the end each party will get a number of seats roughly proportional to its percentage of the vote. Thresholds, that is, rules that exclude parties that do not reach a determined percentage of the vote, can be envisaged. Germany’s five per cent threshold is often cited as a reasonable benchmark, though smaller parties may have a different view of reasonable.

Mixed systems of that kind have been popular outside Canada throughout the 1990s. What could we expect if we went in that direction? It is a common observation that in debates over electoral reform both sides tend to use every possible argument in order to carry the day. Reformers tend to portray their favourite formula as a cure-all for political problems, thus at times raising elaborate expectations, while their opponents often reply in kind, raising fears that do not stand scrutiny. Advocacy has its rules, but political scientists’ job is to try to help separate the wheat from the chaff. Although in the end I conclude that a German-style system would have more advantages than disadvantages, I hope my study has followed this more sober-minded approach.

I will not repeat in detail all the points it makes, but instead summarize in point form what I believe to be the reasonable and unreasonable fears of opponents, as well as the reasonable and unreasonable expectations raised by proponents. Readers who want more detail on each point are invited to consult the full study (which can be found at

Among what I conclude to be the unreasonable expectations about PR is that it will:

  • lessen party discipline
  • single-handedly resolve the crisis of Canadian federalism
  • end the crisis of confidence in politicians and institutions
  • make politics more consensual

Among what I see as unreasonable fears are that PR will

  • generate two classes of warring MPs
  • result in chronic instability
  • make governance worse
  • lead to the breakdown of democracy.

The following expectations strike me as being reasonable. In my view, PR can be expected to produce:

  • a more representative Parliament
  • more smaller parties, although this depends on the stringency of any thresholds
  • less regional polarization
  • more regionally-balanced party caucuses
  • better representation of various regions in Cabinet
  • no more single-party majority cabinets
  • coalition cabinets that are less durable than present Cabinets
  • weaker Prime Ministers
  • a higher voter turnout

And, finally, I think the two following fears are quite reasonable. PR probably will produce

  • less emphasis on constituency work for list MPs
  • coalitions that are decided after elections by party leaders, rather than governments that, as now, are more or less “directly” elected by the voters.

Electoral reform is not a cure-all. That many people apparently have come to believe it is probably is due to the faulty working of first-past-the-post in recent years, and to disenchantment with the Westminster model in many of the countries where it prevails. Alternative voting is clearly a minimalist reform and may not work as its supporters assume. A German-style mixed system would address most concerns commonly raised by reformers, and might be a step towards a less quarrelsome federation.

Photo: Shutterstock