Has Carney Peaked Too Soon?

On Polling Dead Zones

What would happen if every seat the Liberals are leading the Conservatives by 10% or less went blue?

There’s been some discourse about whether the Liberals are “peaking too soon”, as if the Liberals would be better off to be polling at 40% and not 43%. It’s a stupid concept - you want a lead as big as possible so that when - and I do maintain it’s a when - things fall backwards for you, you’re falling from a big lead to a smaller one, not from a small lead to a tie or a deficit.

One of the problems for Kamala Harris is that she was never able to stretch out a big lead, peaking on average around a 4% lead nationally that eroded back. She never had a huge lead, in part because of her own choices and in part because of circumstances. At no point did I ever write or say anything more confident than some versions of “she’s favoured, but obviously Trump can win and it wouldn’t be surprising at all if he won” for a reason.

Now, Carney’s up big, but this week’s polling surge has barely moved the model. And there’s a reason for it that’s best explained by that hypothetical I gave at the start of the column.

The model right now has the Liberals at 194 seats and the Conservatives at 123. The Liberals have a 22 seat cushion above majority government, but that’s less informative about how close the race is than how much you’d have to move the polls back right to get to a competitive race. And the answer’s simple - a lot.

There are 24 seats in Canada right now where the Liberals have a 10% lead or less over the Conservatives. Of those 24 seats, 13 of them are within 5%, and another 11 between 5% and 10%. Even if the Conservatives won every one of those seats, the Liberals would still have 170 seats, the Conservatives would stall out at 147, and the Liberals would easily govern.

Now, obviously if the Liberal vote fell by 5% and the Conservative vote rose by 5% you’d see effects elsewhere - Elizabeth May, Peter Julien, and Leah Gazan would all hold their seats, and 3 Liberal seats in Quebec would go Bloc - but the broad strokes stay the same. The Liberals would gain seats, the Conservatives would too, the NDP and Bloc would fall, and the Liberals would be the only plausible government. And the reason’s simple - even though the Liberals haven’t been gaining much in terms of seat totals this week as their polls keep chugging upwards, they’re creating big margins in what we think of as competitive seats. And that means the Liberals can survive a lot of polling regression before losing a lot of seats.

Let’s take BC, where the Liberals are a comical 16.6% up on last time in my polling average. I don’t believe that for a second. Not for a singular second do I actually think Carney will got 43% of the vote in BC on election day. But, right now, a 10% swing on margin to the Conservatives would only cost the Liberals two seats there - a Kelowna area seat that just flipped over in today’s model update, and Abottsford, which is only competitive because Michael De Jong is running as an Independent after being screwed out of the Conservative nomination. The next most vulnerable Lib-Con battle in the province? South Surrey White Rock, where the Liberals are up 11%.

In SaskyToba, the least safe Lib-Con battle is Winnipeg West, where the Liberals are up a robust 14%. Quebec would only see Trois Rivières flip under these swing conditions, and for the Conservatives to gain a seat in Atlantic Canada they didn’t win in 2021 would require a 13.5% swing from where we are now.

Where there are a significant number of close races is Ontario, but even then, the seats that aren’t close tell the story better than the ones that are. Nothing in Brampton is within 16% even with a significant modelled effect for minority voters moving right, the closest Mississauga is still Liberal by 9.7%, Cambridge is a 11% race, both Burlingtons are 15% leads or bigger, the more competitive Oakville is a 9% race, the new Milton East-Halton Hills South seat that voted Liberal by 0.9% in 2021 is LPC +7.6%. All of the traditional battleground areas, the ones that decide government, aren’t close.

Even something like Kitchener Conestoga, which the Liberals flipped in 2019 and held in 2021 by tiny margins both times, is a 7% race right now. These aren’t close races, which means that the Conservatives need close to 10% reversion in the polling to get close to a draw in seats, plus pray that the Liberals tank and the Bloc rise in Quebec.

The reason the model top lines have barely moved this week is that the Liberals have been in a dead zone, as it were - they essentially hit the ceiling of plausible gains, and there’s a big gap between those and the next tier of targets. What that means practically is that the Liberal vote has gotten a lot less efficient, because 43% of the vote yields pretty much the same number of seats as 40%, but at 43% those seats are a hell of a lot safer.

The theoretical problems for the Liberals are real, and Carney’s fuckups around the Polytechnique shooting yesterday are a good reminder for Liberals that things are by no means rock solid for them. But that can be true at the same time as another point: the Liberals‘ polling can do down a considerable ways before they lose their majority. They’re not impervious to political gravity or reality, but they’re also remarkably well positioned to sustain some falling polls without losing many seats off their projections.