Introduction Edit

Canada's current 'first past the post' electoral system, adopted from the UK, has been used since Confederation. It has the virtue of being simple, and it has the force of long tradition. However, Canadians would be more democratically represented by another system that more accurately reflects voter preferences.

The Issue Edit

Citizens' voting preferences should determine who forms government. The electoral system should be designed to determine citizens' voting preferences as accurately as possible. Under Canada's current electoral system, it is possible for a party to obtain control of the government with less than the majority of the popular vote, because many seats are won by candidates who win a plurality (but not a majority) of votes in their single-member constituency. Moreover, many voters end up voting "strategically" for a candidate who is not their first choice in an effort to defeat an even more unpalatable candidate. This distorts voter intentions and results in unrepresentative Parliaments.

Another problem with first-past-the-post is that it exacerbates regional tensions by causing regional overrepresentations of particular Parties in Parliament. For example, most Albertans support the Conservative Party of Canada (CPC); under first-past-the-post, all or nearly all MPs from Alberta are Conservatives, despite the fact that the Liberals generally win at least 15% of the votes in Alberta, and about one-third of Albertans vote for Parties other than the CPC. First-past-the-post means there are very few or no Liberal MPs from Alberta. As a result, no Albertan voice is at the Cabinet table when there is a Liberal government; hence contributing to 'western alienation'. Conversely, because of first-past-the-post, the Conservative party is underrepresented in Canada's three largest cities. Similarly, the Bloc Quebecois is significantly overrepresented in Parliament compared to its share of the popular vote in Quebec, giving unearned momentum to the separatist cause.

Yet another serious problem with first-past-the-post is that Parties with significant, yet geographically dispersed national or regional support are unable to send MPs to Parliament, effectively disenfranchising all voters who identify most closely with those Parties. In many regions, for example, the Green Party garners more than 5% of the vote; yet it has no MPs in Parliament. Similarly, the New Democratic Party regularly sends far fewer MPs to Ottawa than it would under an electoral system that represents Parties according to their share of the popular vote.

These distortions of the people's electoral will are unnecessary and unacceptable. Strategic voting needs to be rendered obsolete, and the number of MPs a Party sends to Ottawa from each region should directly reflect that Party's regional level of voter support.

Electoral reforms can achieve these goals.

Alternatives Edit

There are two basic methods which may overcome the issues the current system suffers from.

(Party-) Proportional Systems Edit

In their most basic form, proportional systems have no local seats. Instead, a party is simply given seats in Parliament in proportion to how many votes they received federally. Seats go to the party not the candidate, and the party allocates them to its own candidates in an order determined by the party, not by the voters themselves.

So, if the NDP received 20% of the vote, they would receive 20% of the seats. These seats would be chosen from Party lists, with the candidates ranked in order of priority; thus, if the NDP list included 308 candidates, but the NDP's share of the national vote assigned it 61 seats, the first 61 names on the NDP's ranked list of candidates would be elected.

Where MPs are elected according to their order of priority in Party list systems, each Party's internal method for assigning rankings to candidates is enormously important. To ensure grassroots democratic control, the Canada Elections Act would need to specify detailed procedures to ensure Party lists are arrived at through democratic, transparent and accountable election processes which all members of a Party are given an opportunity to vote in. This is extremely unlikely in today's environment where effectively citizens have no control over internal party process, parties themselves can't be sued (at least not directly), and party constitutions have no legal status. Canadians have seen popular MPs such as Sheila Copps and Garth Turner ejected by their parties, and would be very unlikely to hand the power to exclude whistleblowers from parties, leaving the public no power to elect independents or support MPs unpopular with leaders.

A modification of a national proportional representation system more suitable to our enormous and diverse country might feature regionally organized Party lists rather than national Party lists. However, in November 2005 in PEI (a very small province of under 150,000 people) a system that allocated just over a third of seats by such lists, failed badly with a 38 per cent referendum result.

In order to ensure that Canadians would still have a local MP to turn to as their intermediary with Ottawa, the MPs from each Region could be assigned to local constituencies (Ridings), even though they were elected from Party lists rather than constituencies. Again, the party list would have to determine which candidate got to choose which riding first, since there'd be no information on which candidates were the most popular with the public.

Majoritorian Systems Edit

Majoritorian systems do not attempt to ensure that the parliament is an exact reflection of the nationwide popular vote. Rather, they ensure that the candidate in each riding was chosen by a majority of people within that riding. Single-winner Single Transferable Voting (STV) or Instant Runoff Voting, Condorcet, and Approval voting are different methods of achieving this.

Instant Runoff Voting (IRV) is used in Liberal Party of Canada nomination battles when three or more people are vying to become the party's candidate in a federal election, because it is the fairest and most efficient way to select winners in a single-member constituency election process.

  • Is it? I thought it was just a runoff

With IRV, voters list candidates in order of preference. For example, in a Riding with seven candidates, a voter may like three of them enough to give them some support. She would list these three in order of preference: 1, 2, 3. Winners are determined by eliminating the least-preferred candidates progressively, and re-assigning their votes by re-examining the order of preference on ballots until one winner emerges who necessarily enjoys the support of a majority of the constituency's voters (which is why it is a 'majoritarian' voting system).

Combined Systems Edit

Various systems exist which attempt to combine the best elements of constituency-based and Party-proportional electoral systems. Such 'mixed-member' systems are very likely the best solution for Canada, since they encompass the virtue of having Members of Parliament responsible to specific Ridings in our vast country, and of a Parliament which reflects the overall national and regional distribution of voter support amongst Parties.

The number of seats allocated by direct votes in the district/region and allocated by party-proportionality varies. Some systems propose a 50-50 split, but in PEI in 2005 a system with just over a third (10 of 27) seats allocated by party lists was soundly defeated. Since only a few seats are actually required to make up for disproportional results in the nearly-party-proportional STV system, the number of such party-allocated seats could be as low as 10 per cent or even 5 per cent.

Also, gerrymandering is a common problem in all electoral systems, but especially in North America where parties often manipulate processes that set district borders so as to favour their own candidates, or maneuver situations where ridings are merged or split to disempower locally popular candidates. Grassroots organizing is difficult if municipal, provincial and federal district borders don't match (a special problem again in North America with three levels of government) and people are being told in nearly every election that they are now "part of" a community that they weren't last time. Bioregional democracy advocates also argue that splitting sensitive features of the environment such as watersheds or rich agricultural areas, results in poor representation of concerns regarding the commons. Some argue for a "commons sector" to represent these outside the electoral system, others for borders that more closely match nature's and which are set by ecologists and urban planners, not by politicians.

Because the number of people in each such district may vary widely, a flexible combined multi-member system is required. Several of these bioregional multi-member district systems exist.

BxxV+C+P (Hubley model) Edit

Two examples that accomodate fixed bioregional borders and solve the problem of party proportionality while avoiding party lists (but could accomodate them) are:

  1. the BSTV+C+P (bioregional single transferable vote plus circuits plus party-proportionality) proposal based on work by Julian West and Craig Hubley, and
  2. the B5AV+C+P (bioregional five-approval-vote plus circuits plus party- proportionality) which replaces the transferable or ranked vote with five votes or points that can be used to approve or disapprove a candidate, an innovation proposed by Craig Hubley and Martin Willison.

Both of these systems exploit that fact that fixing permanent bioregional boundaries will result in numbers of members that are not whole numbers, therefore an extra .1 to .9 seats will be owing to the district after whole numbers of members are elected. Adding these up and assuming an average of .5 such missing members per district, a system with 20 districts would have ten missing members. This is enough to make up for party proportionality problems - while the electorate as a whole determines the proportionality, the votes of the districts "most cheated" will count disproportionately (more for districts missing .9 members, less for those missing only .1) in determining which members of the under-represented parties go to the legislature to represent their party. If districts with .0 missing members (whole numbers) exist, their votes count for overall proportionality but do not determine which members is elected. There is no party list, members go to the legislature in the order of popular vote modified by the weighting that makes up for fixed borders.

Neither system absolutely guarantees that passing a certain threshold of votes set in advance will result in seats for small parties - though this could be accomodated by allocating at least one seat for the leader of any party that breaks through such a threshold. If there is no such rule, then all parties are at least guaranteed that no party with lesser support gains a party-proportional seat before a party that has more support, and accordingly that for any given election a minimum threshold can be observed after the fact. This guarantee is enough to ensure that no one is favoured.

Both systems also permit districts to have 1.0 members to represent remote areas with small populations, another special problem of North America. Since STV degrades very gracefully to instant runoff in the single-winner case, this amounts to an instant runoff ballot for the member, and all votes counting in the party-proportional count. Depending on the weighting of votes in .0 districts which are missing no members, there would be very little chance of defeated candidates in such a 1.0 district getting a party-proportional seat, even if they came second and were the most popular politician of that party in the whole election. Therefore the chance of a party leader choosing to run in a remote region is reduced. For these reasons, it is sometimes suggested that the most popular politician of a party, or the leader, should get the first party-proportional seat allocated no matter what the weighting. This would however reduce the number of seats available to correct proportionality in general, and so is probably only practical in large provinces with a great many seats.

Complexity Edit

Some people argue that combined systems are too complex for the average voter to understand. However, the average voter doesn't understand the current system either. Many are understandably quite angry at results that deny representation to their party that doesn't match the reported popular vote, or demands that they "strategic vote" to avoid a bad result of vote-splitting. People seem to prefer that they simply express their desires and let the system work out how exactly to do that.

In practice, complex electoral systems have proven perfectly acceptable in Australia, France, Germany, Israel and other countries. In BC in 2005, the quite complex BC-STV system was supported by 57.4 per cent of voters and 77 of 79 districts, missing very narrowly in the other two. Only overly strict approval thresholds (which did not allow a near-unanimous support by the districts to make up for a missing 2.5 per cent of the popular support overall) prevented this from actually taking effect in 2005. In fact, Gordon Campbell's government gained a false majority with less than 50 per cent of the popular vote, meaning, many more people believe that his government should not have that power, than voted to give it that power. This is uncomfortable and leads to authority challenges against the government, and probably ultimately undermines its authority more than a complex system that all the other parties accept.

MMP (Mixed Member Parliamentary representation) Edit

MMP (Mixed Member Parliamentary representation) elects some Members of Parliament from geographic constituencies, and some from regionally organized Party candidate lists. Each elector fills in two ballots: one to elect their local constituency representative, one to choose a Party preference.

Two-ballot MMP may be truly proportional, or it may be 'parallel'. Truly proportional MMP is used in Germany. Under the German system, half of the seats in the Bundestag are elected from single-member constituencies; the other half is derived from Party lists. The Party list seats are distributed with the mathematical objective of adding seats to each Party's delegation to the Bundestag until the total number matches the Party's overall level of national support as closely as possible (as measured by the Party preference ballot). To prevent the Bundestag from splintering into a multitude of small Parties, a cutoff of 5% is set: any Party with less than 5% of the popular vote gets no seats in the Bundestag.

In Germany, the method used to elect single-member constituency representatives is 'first past the post.' However, it would be straightforward to design a reformed Canadian electoral system to elect single-member constituency MPs by a majoritarian system instead, such as Instant Runoff Voting, in order to eliminate the incentives for hold-your-nose 'strategic' voting imposed by first-past-the-post.

Parallel MMP systems count constituency seats and Party list seats separately. For example, imagine a Canadian Parliament with 308 single-member constituency seats and 154 Party list seats. The constituency seats should be elected by majoritarian IRV. The 154 Party list seats would be elected by straight proportional representation (with a 5% minimum to prevent splintering).

Under a parallel mixed-member system, a Party that was unable to win a single constituency election, yet gained 10% of the votes on the Party list ballot, would send 15 MPs to Parliament (10% of 154 seats). In contrast, under the German-style system of fully proportional representation, the same Party would send 46 MPs to Ottawa (10% of the full 462-seat Parliament) even though it didn't win a single constituency election.

In November of 2005, a referendum was held in the province of Prince Edward Island, asking voters if they would prefer to switch to a mixed-member system. It was defeated, with only the city of Charlottetown in favour of the switch. This decision may be attributable to the fact that P.E.I., with a population of 138,000, is far too small a jurisdiction to require the complexity of MMP.

STV (Single Transferable Vote) Edit

STV (Single Transferable Vote) works similiarly to a majoritorian system, with the difference that each geographic constituency sends multiple MPs to parliment. (These constituencies or Ridings would perforce have to be much larger than current Ridings.) It works on a preferential voting basis. As with IRV, the basic premise is that each voter ranks the given candidates on the ballot from their most preferred, to their least preferred. These votes are then redistributed in several rounds of counting, as candidates who garner the least number of first-choice votes are "dropped" and the ballots which marked these candidates as the first choice are re-examined to count the second choice instead, and so on in several rounds of counting until winners are determined. Thus STV takes into consideration the voters' second, third, and perhaps even fourth choices in selecting winners.

In May of 2005, on the day of the British Columbia general election, voters were asked if they wished to switch to an STV electoral system as opposed to the current single member plurality. The threshold for success was undemocratically high: 60% overall voter support, plus a majority of voters in 60% of the 79 constituencies had to be in favour of the change. Although only 2 of the 79 ridings voted against STV, and 57.7% of voters favoured STV, the referendum failed to garner 60% in favour, and the STV proposal was defeated. Premier Gordon Campbell, acknowledging the close result, has decided to pursue further studies on electoral reform, which may conclude with yet another referendum in the next provincial election, in 2009.

  • Unless somebody know something I don't, there will be a referendum in 2008

This policy is a stub. Please help us by expanding it.