Minority Math is Hard on the Liberals
September 28th, 2015 | 17 Comments »
---- 2 ±±±± 1 ±±±± 0 ±±±± 1 ±±±± 2 ++++
The Liberals have run a strong and energetic campaign thus far under leader Justin Trudeau, so it seems unfair how hard it will be for them to find enough seats to claim the head of a minority government.
Whether you're following one of the popular poll aggregator and seat projectors, or just working from your own spreadsheet and rolodex, the Liberals are stuck stubbornly in third place in the seat counts, even as they've topped or come second in the daily tracking or polling averages. One political pro who has run the simulations says they'd need a 6-point lead before they could claim the largest number of seats.
Why is that? A number of factors come into play:
- While the Liberals are up 10-12 points from their 2011 vote-share of 18.9% nationally, outside Quebec that's mainly come out of the hide of the Conservatives, putting more ROC seats into play across the country for the NDP. Counterintuitive, if your focus has been on Ontario to date in this campaign, but true nonetheless. The benefits accrue to the orange team mainly in BC, but a smattering as well across the Prairies, and in Northern and Southwestern Ontario. In Quebec, the Liberal gains in vote-share as of Sunday night had come equally out of the Bloc and the NDP.
- The Liberals are only leading regionally in Atlantic Canada, and (mostly) in Ontario, though they have some other concentrated pockets of support: in anglo- and allo-Quebec ridings, Winnipeg, and upscale Vancouver. Even in Ontario, Liberal support is concentrated in the 905, 416, Halton-Peel, and the National Capital Regions. There are plenty of seats to be won for them there, to be sure, but not enough to get over the minority government threshold of 120 or so, especially when they are shut out of large swaths of rural Ontario.
- The Liberals are not competitive in over half of the 338 seats in the new House of Commons. Even if they kept every current seat, and completely ran the table in i) NDP-Liberal races, ii) Conservative-Liberal races, and iii) three-way races, they would still fall well short of the magic 170 to secure even a slim majority.
- By contrast, were everything to go their ways, either the Conservatives or NDP could each conceivably be competitive in around 60% of the seats, albeit not at the same time.
- Racking up larger and larger margins in the Atlantic, anglo- and allo-Quebec ridings, the 905, north Toronto, and Halton-Peel doesn't win the Liberals more seats. By the same token, even at 36% in Quebec with a split opposition, the NDP could still expect to leave the province with the majority of seats. In 2008, for example, the Bloc won two-thirds of the seats in Quebec with just 38% of the vote (against 24L, 22L, 12N). Three years later it won 4 seats with 24%. As of Sunday, it stood at 16.4%.
To help see this pattern, let's first remind ourselves of the parties' performance in 2011. I've grouped the Prairies to be comparable with the CTV Nanos Daily Tracking, and included the count of nominal seat wins based on the new (338-riding) boundaries.
2011 General Election Performance
[seat counts using nominal wins on new boundaries]
| 2011 Vote% [seats/338] |
N60 | BC | Pra | ON | QC | Atl | TOT |
|---|---|---|---|---|---|---|---|
| Cons | 36.3% [2] |
45.6% [28] |
62.8% [55] |
44.4% [83] |
16.5% [5] |
37.9% [15] |
39.6% [188] |
| NDP | 27.8% [1] |
32.5% [11] |
22.0% [6] |
25.6% [24] |
42.9% [61] |
29.5% [6] |
30.7% [109] |
| Lib | 26.4% |
13.4% [2] |
10.8% [1] |
25.3% [14] |
14.2% [8] |
29.3% [11] |
18.9% [36] |
| Grn | 9.3% |
7.7% [1] |
4.4% |
3.8% |
2.1% |
3.0% |
3.9% [1] |
| BQ | 23.5% [4] |
6.1% [4] |
And then, let's look at the CTV Nanos tracking as of Sunday, with the changes from 2011 in brackets below.
CTV Nanos Tracking Poll of Federal Vote Intention, Showing Change since 2011
(CATI, n=1200, September 24-26, 2015)
| Nanos/CTV Sep 24-26, 2015 (chg since 2011) |
N60 | BC | Pra | ON | QC | Atl | TOT |
|---|---|---|---|---|---|---|---|
| Cons | – | 32.5% (-13.1) |
46.7% (-16.1) |
37.6% (-6.8) |
18.2% (+1.7) |
13.2% (-24.7) |
31.5% (-8.1) |
| NDP | – | 31.8% (-0.7) |
21.3% (-0.7) |
23.5% (-2.1) |
35.7% (-7.2) |
26.9% (-2.6) |
27.6% (-3.1) |
| Lib | – | 27.8% (+14.4) |
26.1% (+15.3) |
35.6% (+10.3) |
27.9% (+13.7) |
54.4% (+25.1) |
32.5% (+13.6) |
| Grn | – | 7.8% (+0.1) |
3.0% (-1.4) |
3.3% (-0.5) |
1.8% (-0.3) |
4.7% (+1.7) |
3.7% (-0.2) |
| BQ | 16.4% (-7.1) |
4.0% (-2.1) |
To help sort out the range of seat possibilities, I rated each riding for its range of outcomes, current guestimated winner, and noted the nominal winner from 2011 on the new boundaries.
That yielded seven groups of ridings. Imagine a triangle with one group on each point (core and strongly leaning seats for each party), and one group on each side (the two-way contests between each pair of parties), with the seventh group in the middle (the three-way contests).
If you completed the exercise, you might quibble with the categorization of ridings here and there, but I suspect would arrive at a similar configuration. Remember I was not trying to precisely determine a seat projection for today; just group the ridings by range of outcomes for this election. This yielded the following groups:
- Core and strongly leaning Conservative seats – 70 (42 core and 28 strongly leaning) as follows: 2-NB, 2-QC, 22-ON, 7-MB, 8-SK, 21-AB, 8-BC
- Core and strongish leaning NDP seats – 63 (12 core, 3 based on incumbent strength, and 48 strongly or likely leaning) as follows: 1-NS, 46-QC, 6-ON, 1-MB-, 1-SK, 1-AB, 4-BC. I'm hedging my bets slightly here, as several pollsters were reporting overnight that their morning numbers would show a significant shift, which I take to be some movement in Quebec. I guess we'll see.
- Core and strongly leaning Liberal seats – 18 (15 core, 1 based on incumbent strength, and 2 leaning) as follows: 8-Atl, 3-QC, 5-ON, 1-SK, 1-BC
- Conservative-Liberal contests – 49 (46 currently held by the Conservatives, 3 by the Liberals) as follows: 3-NS, 1-NB, 1-QC, 39-ON, 1-MB, 1-AB, 3-BC
- Conservative-NDP contests – 44 (32 currently held by the Conservatives, 12 by the NDP) as follows: 11-QC, 8-ON, 1-MB, 3-SK, 5-AB, 16-BC
- NDP-Liberal contests – 44 (29 currently held by the NDP, 13 by the Liberals, and 2 by the Conservatives) as follows: 7-Atl, 13-QC, 18-ON, 2-MB, 3-BC, 1-N60
- Three-way contests – 45 (36 currently held by the Conservatives, 5 by the NDP, and 4 by the Liberals) as follows: 7-Atl, 23-ON, 2-MB, 1-SK, 5-AB, 5-BC, 2-N60
Five other ridings are left over:
- BQ-Cons contest – Bécancour—Nicolet—Saurel, QC
- NDP-BQ contest – La Pointe-de-l'Île, QC
- Cons-Ind contest – St. Albert—Edmonton, AB
- NDP-Grn contest – Victoria, BC
- Green Party leader Elizabeth May incumbency seat – Saanich—Gulf Islands, BC
Now suppose each party ran the table in terms of holding their core and strongly leaning seats, and then winning all of the two-way contests they were involved in and all of the three-way contests. That would yield the following theoretical ceilings for each party:
- Conservatives: 70 core and strongly leading + 49 Cons-Lib + 44 Cons-NDP + 45 three-way contests = 208
- NDP: 63 core and strong leading (unless the Quebec numbers massively moved last night) + 44 Cons-NDP + 44 NDP-Lib + 45 3-way = 196
- Liberals: 18 core and strongly leading + 49 Cons-Lib + 44 NDP-Lib + 45 3-way = 156
Next, consider the share of seats if each party were to win an equal share of all the contests it was involved with:
- Conservatives: 70 core and strongly leading + 25 Cons-Lib + 22 Cons-NDP + 15 three-way contests = 132
- NDP: 63 core and strong leading + 22 Cons-NDP + 22 NDP-Lib + 15 3-way = 122
- Liberals: 18 core and strongly leading + 24 Cons-Lib + 22 NDP-Lib + 15 3-way = 79
- Five other seats: 2-NDP, 1 or 2-Cons, 1-Grn, 0 or 1 BQ
From there, everything is a zero-sum game. But, so long as the NDP can maintain a large bloc of seats in Quebec as the anchor of its caucus, it has a much better chance of winning more seats than the Liberals, in spite of a third place finish in Ontario. It's the difference between starting off with 60-some and 20-some seats in your column.
In terms of Quebec's 78 seats, Sunday night no-one was crediting the Conservatives with more than 8-12, nor the Liberals realistically with more than 12-14 either. I discount the projected Liberal wins in the Gaspesie and Bas-St-Laurent, given they were based on 2011 results from a strong local candidate
We'll consider each group of seats in subsequent blogposts, and take a look at where the Leaders have been concentrating their attention.
