Potvin's Argument of Bias

In an opinion piece published August 29th in the Vancouver Courier, Kevin Potvin argues that STV will favour the Liberal party. His argument is based on the assumptions that support for the Liberals is highest in the rural parts of the province and that smaller districts favour the Liberals; since the STV districts in the rural parts of the province are on average smaller than those in the more urban parts of the province, he concludes that STV will produce a structural bias in favour of the Liberals.

Counter-Argument

Fortunately for BC voters, both his assumptions and his conclusions are incorrect. First, there is no significant bias towards the Liberals in the rural parts of the province. Second, smaller districts do not favour the Liberals. Finally, when we take a detailed look at how the seat count would likely shake out under STV, we find that STV will do an excellent job of balancing out the strong bias in favour of the Liberals that exists under the current First Past the Post system. We tackle each issue in turn below:

1. No significant bias towards Liberals in rural parts of province.

For the purpose of this analysis, we consider the 23 ridings in the BC interior ('beyond Hope') that existed for the 2005 election. There were 424,000 votes cast in these 23 ridings. The Liberals received 45.8% of these votes, the NDP 40.9% and the Greens 8.2%. Provincewide, the Liberals received 45.8%, the NDP 41.5% and the Greens 9.2% [see Elections BC for details]. These numbers are virtually identical.

2. Smaller districts do not favour the Liberals.

Potvin argues that smaller districts will favour the leading party, but this represents a misunderstanding of how STV works. In a two seat district, a candidate requires 33% of the vote to win a seat. In the Peace River area, for example, the Liberals took 59% of the vote to the NDP's 30% and the Green's 8%. The Liberals can clearly win one seat easily, but they will have only 26% of the vote to put towards the second seat. If the NDP can attract 3% of the Green's 8% share of the vote, they will have enough to win the seat. On the other hand, if the Greens and the other 3% of the voters transfer their votes to the Liberals, the Liberals could end up with 70% of the vote and will win the second seat. In both cases, this is a fair outcome. In the Northwest, the Liberals had 44% of the vote vs 47% for the NDP, and each party would win one seat.

3. Detailed analysis of outcome with STV: Liberals 11, NDP 10

In 2005, there were 23 ridings. The Liberals won 15 (65% of them) on 46% of the vote, while the NDP won 8 (35%) on 41% of the vote.

Under the redistribution proposed by the Electoral Boundaries Commission, the Interior would have 21 seats arranged into 7 STV districts containing 2 to 4 seats each. The table below shows how we have roughly assigned results from the previous 23 ridings to each of the proposed new districts. There will be some discrepancies around the edges because the boundaries no longer match up precisely, but the general trends should be apparent.

Correspondences between proposed STV districts and 2005 ridings

• Northwest: Bulkley Valley - Stikine, North Coast, Skeena
• Northeast: Peace River North, Peace River South
• North Central: Prince George - Mount Robson, North, Omineca
• Cariboo-Thompson: Cariboo North & South, Kamloops, Kamloops-North Thompson, Yale-Lillooet
• Columbia-Kootenay: Columbia River-Revelstoke, East Kootenays, Nelson/Creston
• Okanagan North: Shuswap, Okanagan-Vernon, Kelowna Lake Country, Kelowna Mission
• Okanagan South: Okanagan Westside, Penticton-Okanagan Valley, West Kootenays-Boundary

Vote shares in STV districts and estimates of seats won

In these seven STV districts, the Liberals, NDP and Greens won the following percentage of the vote (equivalent quotas shown in brackets):

• NW (2 seats): 44/47/6% (1.3/1.4/0.2) Likely seats: 1/1/0. 2005 seats: 1/2/0
• NE (2 seats): 59/30/8% (1.8/0.9/0.3) Likely seats: 1/1/0. 2005 seats: 2/0/0
• NC (3 seats): 48/36/8% (1.9/1.5/0.3) Likely seats: 2/1/0. 2005 seats: 3/0/0
• CT (4 seats): 46/44/7% (2.3/2.2/0.4) Likely seats: 2/2/0. 2005 seats: 2/3/0
• CK (3 seats): 37/52/10% (1.5/2.1/0.4) Likely seats: 1/2/0. 2005 seats: 1/2/0
• ON (4 seats): 49/33/9% (2.4/1.6/0.5) Likely seats: 3/1/0*. 2005 seats: 4/0/0
• OS (3 seats): 45/42/9% (1.8/1.7/0.4) Likely seats: 1/2/0**. 2005 seats: 2/1/0

Totals: Liberals 11, NDP 10, Greens 0

* it is likely that the fourth seat in Okanagan North would go to the Liberals because of the number of votes that went to right-leaning independents and smaller parties.

** this result is debateable, but if the Green votes go predominantly towards the NDP, they would take the last seat. This seat would certainly be competitive.

Government AND opposition in all regions

Note that under FPTP, the Northeast (Peace River area), the Prince George area and the Okanagan Valley all went 100% to the Liberals (the single NDP seat in the south Okanagan is technically the West Kootenays-Boundary seat). Under STV, the NDP would likely win at least one seat in each region, which is a much more representative result. In addition, even if an NDP candidate did win under FPTP, they would technically only represent their own riding and NDP supporters in other ridings would still be without representation; under STV, these opposition MLAs would represent NDP supporters throughout the district.

Effect of 'exhausted' ballots

One final note: Brian Bush asked in a letter to the editor of the Vancouver Courier published on Friday, Sep 21, 2007 what the effect would be if a signficant number of voters only mark one choice on their ballot. In most such cases, where the indicated candidate does end up winning, the voter's ballot will count towards their election. If such a candidate is eliminated, there are no indications on the ballot as to which of the remaining candidates the voter prefers, so elections officials will have no choice but to place the ballot aside. Voters will need to be educated to list preferences for as many candidates as would be necessary to be reasonably sure that one of the candidates on the list will be elected.

Conclusions:

• Under FPTP, the Liberals win a disproportionate 65% of the interior seats on 46% of the vote, while the NDP win 35% on 41%.
• Under STV, the Liberals will win 11 of 21 seats (52%) while the NDP will win 10 seats (48%). This is closely proportionate to their relative vote shares.
• The Greens do not yet have sufficient strength to win a seat outright, although they might win sufficient vote transfers in Okanagan North to be competitive for the last seat. In any case, their supporters will likely have the deciding vote in several contests, so will have an opportunity to elect candidates who are generally supportive of their policies.