As the country enters into week 5 of the election campaign, it appears as though a number of trend lines are emerging that might carry through to voting day.  The English language debate came and went, by most analysts’ reckoning, without a knockout punch.  It was the French debate that might have been more decisive, directly at the expense of the Conservatives, but ultimately at the expense of the Liberals.

Several polls since that debate indicated that the Bloc had made up some ground in Quebec, with two very recent polls showing the race in Quebec essentially tied between them and the Liberals.  This shift in popular support came largely at the expense of the Conservatives, who were widely seen to be the losers in French language debate. They had less to lose, however, since the most optimistic projections showed maybe a dozen or so ridings in Quebec going blue.  The Liberals have largely kept their support, maybe a few points less in some polls, but largely intact.  The problem they have is in losing any amount of Quebec support, to one main rival (i.e. the Bloc), rather than a vote split, they run into the math of voting inefficiency, which previously beset the other parties far more. Any gains from the Bloc are likely to be sufficiently disbursed around the province to put potentially 2 dozen seats into serious contention (which my model reflects); the Liberals meanwhile remain strongest in the western part of Montreal and are very dependent on voting splits elsewhere to win seats.

Ironically, the Conservatives losing support in Quebec might actually help them form a government, possibly more so than if they had held onto or expanded their support.  If Bloc gains at their expense (again, in votes, not so much in seats since they weren’t going to win that many to begin with) – that hurts the Liberals and quite possibly edges the Conservatives slightly ahead in the seat count.

It is worth noting that a lot of time remains between now and the election, and the dynamics of a tight horse race seem likely to continue during the remainder of the campaign.  A subsequent french language debate might see Trudeau successfully reverse his fortunes in tampering down any Bloc resurgence; alternatively a foreboding sense on the part of left-of-centre voters fearing a Conservative government might prompt them to coalesce behind the strongest left-of-centre candidate in marginal ridings at the eleventh hour, and push the Liberals/NDP/Greens past the 170 mark.

The road to 24 Sussex appears to run through Quebec, having the most ridings in play with fairly tight margins, it would not take too much to tip the balance towards or away from the Liberal incumbents. A strong performance by Andrew Scheer in the next French debate, hurting the Bloc, may just turn out to be Justin Trudeau’s best hope for a second term.

Conservative – 32.73% and 141 Seats

Liberal – 31.27% and 132 Seats

Bloc Quebecois – 7.2% and 33 Seats

NDP – 15.87% and 28 Seats

Green – 9.7% and 2 Seats

plus M. Bernier for the PPC in Beauce, Ms. Wilson-Raybould, IND, in Vancouver Granville

I’m still in the process of fine tuning my model, basically things have changed very little in the last week, with most polling showing a slight Conservative lead in the popular vote and a small Liberal advantage in terms of seats. This week the Liberal seat count was revised slightly downward, in trying to model the most realistic seat outcome with a 2 point popular vote deficit. I still think York Region above Toronto might have 2-3 more red ridings, which netting against maybe 1-2 seats less in the Lac St. Jean area might push the Liberals just above the 150 seat mark. Also different this week, a potential Liberal hold in Calgary, and likewise for the NDP in Rimouski. A number of ridings are truly on the knife’s edge, places such as Kenora, St. Catherine’s, Kitchener South Hespeler, Burnaby-North Seymour (all coded blue this week); these could almost as easily be shaded another colour, and influence who forms the next government. At this point, this is looking very similar to another previous election in October, this one held the day before Hallowe’en back in 1972, when another Trudeau, after having won a resounding majority 4 years prior, barely scrapped ahead in seats (2 ahead of the Progressive Conservatives) and clung to power over the next 20 months propped up by the NDP.

Liberals – 32.32% – 149 Seats

Conserv. – 34.91% – 141 Seats

NDP – 14.74% – 27 Seats

Bloc – 5.03% – 16 Seats

Green – 9.44% – 3 Seats

People’s – 2.75% – 1 Seat

Other – 0.81% – 1 Seat

The second week in the Canadian election campaign might have been overshadowed by goings-on in London and Washington. Indeed, the campaign needle did not move much in Canada, with apparently minimal fallout over the blackface/brownface revelations.

The Liberals remain the party to beat in Central Canada, which will prove critical in the overall election’s outcome. While my model had some adjustments, slightly trending more Liberal as last week I tried to factor in some fallout over the revelations against the Prime Minister, it essentially remains a locked race. At most 30-40 ridings remain in truly in play at this point, and unless the polls move markedly away from the Liberals it is likely not to change.

I keep thinking it will be a close enough race on election night that one commentator will wryly observe if Jody Wilson-Raybould (re-elected as an Independent) will rejoin the Liberals to keep them in power – that it will come down to one seat. Likely, however, the Liberals will be able to win a sufficiently large plurality and join forces with the NDP and Greens to maintain their hold on government.

My model is still needing some fine tuning, I believe, as I think York Region will end up being more bell weather and Liberal red than this map portrays, likewise I am thinking that around the Lac St. Jean region that there might be lighter shades of blue and Bloc sea-green than Liberal red.

The graph below shows that sheer voter efficiency in Central Canada is vital to the Liberals retaining their hold on power. As shown below, the overall percentage of the vote for victorious candidates shows that the Liberals have even a handful of seats won by less than 30% (many of them in Quebec); while few wins are over 60%. By contrast, the Conservatives are showing no wins with a candidate getting less than 30%, notably less MPs being elected with between 30-40% of the vote compared to the Liberals, and on the upper end around two dozen MPs being elected by over 60% of the vote. Therefore, even if the Conservatives edge the Liberals out in the popular vote, which many polls as well as my model suggest, it will be less efficient with wider, deeper blue margins in Western Canada, while the Liberals win on lighter shades of red in Central Canada.

At this point, my model overall is saying

Liberals – 33% and 159 Seats

Conservatives – 35.4% and 134 Seats

NDP – 14% and 26 Seats

Bloc Quebecois – 5.1% and 14 Seats

Greens – 9% and 3 Seats

People’s Party – 2% and 1 Seat (M. Bernier’s in Beauce)

and Ms. Wilson-Raybould, in Vancouver Granville, as an Independent M.P.

This is my first attempt at predicting the upcoming election … some fine tuning and revisions are probably in order for the duration of the campaign … right off the bat I am thinking about revising some of my predictions around York Region in Ontario and the Lac St. Jean region in Quebec – i.e. the Conservatives and Liberals might swap colours on the map.

I am trying to keep a colour scheme such that the sea-green Bloc seats can be distinguished from the bluer Conservative ones.

At this point, trying to build into my model the potential fall out over the brownface incidents, I am showing:

Liberal – 32.1% and 145 Seats

Conservative – 36% and 144 Seats

NDP – 14% and 25 Seats

Bloc – 5.1% (22.3% in Quebec) and 20 Seats

Green – 9% and 3 Seats

People’s Party of Canada – 2.3% and 1 Seat (Bernier’s)

Other – 1.2% and 1 Seat (Jody Wilson-Raybould)

The first week of Canada’s 2019 election campaign has already gone by, and there really appears to be no essential change in the electoral landscape. The Liberals are tied with the Conservatives – most polls showing a slight lead for one or the other party within the margin of error. This is seen as favoring the Liberals, with greater vote efficiency in Central Canada, (smaller margins of victory) whereas the Conservatives can expect more lopsided margins in the western provinces.

This is not to say that things can change, indeed the Liberals, having the edge in projected seats, have the greater risk of losing ground, particularly the seat-rich 905 suburban ring around Toronto which will likely prove critical in the next government’s formation.

Several developments that might have some bearing on the election campaign include the inclusion of Maxime Bernier’s People’s Party in upcoming leaders’ debates, which could adversely impact Conservative vote share (especially in very marginal Ontario seats); the other significant event was today’s revelation that Prime Minister Justin Trudeau had donned a brownface costume while a teacher at a British Columbia private school back in 2001. This could adversely affect the Liberals, who have insinuated that the Conservatives have recruited some questionable candidates and raised the issue of Conservative leader Andrew Scheer’s previous opposition to same sex marriage back in 2005. With this latest revelation, the Liberals could find themselves on the defensive and their attacks on the Conservative’s perceived lack of inclusiveness could fall flat. Should this gain additional traction, it is probably more likely that potential Liberal vote losses would result either from migration to the other left-of-centre parties or depressed voter turnout rather than a ground swell of support to the Conservatives.

The underlying dynamics still strongly support a Liberal win, however, as a majority government in it’s first term has in Canadian political history seldom been denied the chance to continue with a second mandate, even if in a minority position. For all of the recent tarnishing of the Liberal, and more specifically – Trudeau brand, Scheer has thus far not been able to decisively run away with the race on the question on who would make the best Prime Minister. The Conservative brand has not been able to move ahead, in part because of the underlying dynamics to support an incumbent government, and also in part due to the unpopularity of Ontario’s current PC Premier, Doug Ford.

The latest revelations about Trudeau’s brownface soiree nearly two decades ago will undoubtedly make the race more interesting, and potentially much closer, but there is still a lot of campaigning left, and more opportunity for other leaders to make their own gaffes. In Canadian politics many elections have been won not by one’s strength but by one’s opponent’s weaknesses. Nearly five weeks in an election campaign can be an eternity.

Next week we will have available a detailed riding map projecting our winners based on polls, and in subsequent weeks those forecasts will be updated until the week before the election.

Due to a technical glitch on ericvotes.com, we couldn’t update the website with Manitoba predictions, however we did post our pre-election predictions to facebook:

4/6 September 10th … Manitoba General … Pallister wins re-election as NDP gain a few urban seats … PCs 43% and 33 seats; NDP 34% and 21 seats; Liberals 14% and 3 seats; Greens 7 % and no seats; First, Forward and Communists, others, etc. 2% total and no seats

Tenn Can

September 9 at 9:43 PM ·

5/6 … Manitoba … Liberal seats include Burrows; River Heights and St. Boniface

NDP seats … Assiniboia, Brandon East, Concordia, Elmwood, Flin Flon, Fort Gary, Fort Richmond, Fort Rouge, Keewatinook, McPhillips, Notre Dame, Point Douglas, St. James, St. Johns, St. Vital, The Maples, The Pas-Kameesak, Thompson, Tyndall Park, Union Station, Wolseley

Tenn Can

September 9 at 9:51 PM ·

6/6 Manitoba … PC majority seats include Agassiz, Borderland, Brandon West, Dauphin, Dawson Trail, Fort Whyte, Interlake-Gimli, Kildonan-River East, Kirkfield Park, La Verendyre, Lac du Bonnet, Lagimodiere, Lakeside, Midland, Morden-Winkler, Portage La Prairie, Radisson, Red River North, Riding Mountain, Riel, Roblin, Rossmere, Seine River, Selkirk, Southdale, Springfield-Ritchot, Spruce Woods, Steinbach, Swan River, Transcona, Turtle Mountain, Tuxedo, Waverley

The end result was close, a total of 50/57 seats predicted correctly.  The popular vote was a bit wider of the mark, with the PCs winning nearly 50% of the vote, compared to the 43% I had predicted.

Predictions were somewhat complicated by the fact that there was a redrawing of the ridings, with several new ridings in the Winnipeg area where, lacking a transposition of votes from the 2016 election, it was difficult to plug such new ridings as McPhillips in a standard swing model.

The end result was not a surprise – the governing PCs retained a strong majority although they lost some ground in Winnipeg et environs.  Manitoba has historically returned majority governments for at least another term, with the one noted exception being Sterling Lyon’s one term PC majority government, 1977-81, essentially dividing the NDP Premierships of Edward Schreyer and Howard Pawley.  It was further worth noting that this election was held one day prior to the writs being dropped for the Canadian election, with the federal climate suggesting a chill on Trudeaumania 2.0, some retrenchment in federal NDP support, along with a slight shift towards the Greens, while the Conservatives continued a strong hold on the rural areas.  While Manitoba does not always align provincial and federal voting patterns (despite the federal Liberals often winning the popular vote in recent elections, their provincial counterparts have not held power, independent of the Progressives, in nearly a century), in this instance the general direction of the province towards retaining the PCs might translate into some margin gains for the Conservatives at the federal level.