Polling Update: The Conservatives Are In Trouble

On The Latest Round Of Polls

Here’s the basic truth of model making - it’s changes in polls that move the model, not repeated frequency of release. Frank Graves’ work used to move the model quite a lot, but in recent weeks it hasn’t for a simple reason - it’s mostly been constant for a while. So, updating the EKOS line in my polling averages is mostly replacing numbers with minor variations of what was there.

Obviously, if I were to do an average without EKOS it would move things, but what moves the model at this point is pollsters seeing a big change from their last poll. Something like this week’s IRG, which is at a level of Liberal support below many others, did more to help the Liberals in my model than Leger, which is a more Liberal friendly result. The reason is that IRG swung left while Leger was basically status quo.

This all also means it’s not hard to figure out where the model will likely go moving forward. Pallas, which came in with a healthy Liberal lead, swung a couple seats right today, as extremely, extremely, extremely good results for the Liberals last week got replaced with merely extremely good ones this week. Ipsos on Sunday will probably do the same, at least in Ontario, where their 20 point Liberal lead will shrink. (Then again I said that about Angus Reid and they went from a 14% lead to a 16% lead this week, so who knows.)

Monday, I tweeted about this, expecting the Liberals to have a bad week in the model because they were going to lose some (again, maybe even too good to be true) numbers when MQO and Angus Reid and Pallas dropped, and to a lesser extent I was worried about Leger. What happened? They’re at 198 seats right now.

Now, some crunch points - next week’s Angus Reid, Sunday’s Ipsos, maybe a future Pollara or MQO - but by the same token it’s not like the daily trackers have been on fire for the Liberals in recent days. Probably most importantly is the fact that this is another week where the polls are mostly stable, which means that the Conservatives really have 3 more rounds of polling - pre-debate, post-Easter weekend, and then final weekend - to close a 80 seat gap. Or, sorry, a 79 seat gap now. When the main clip coming out of Poilievre’s day is talking to Laura Stone about crowd sizes, I don’t think the campaign’s going well.

So, where do we go from here? It’s increasingly difficult to see how the Conservatives pull this off. I know a bunch of people will think this is a dog bites man analysis from me - it’s the obvious thing I would say because I’m a “snotty Liberal blogger”, but this election is far more difficult for me than, say, Alberta 2023, because in Alberta I knew I wasn’t being biased. The dirty little secret is that the modeller and analyst in me prefers elections where the right are winning because they’re so much easier on my mental sanity. Sure, a few years ago I’d be loving an election like this where I could be a cocky bastard, but I’d like to think I’m a better analyst now than I was then.

The Conservative problem is that they’re going to put on a lot of votes at this election and still probably lose very badly. 2021 PPC voters who like Poilievre will bolster the popular vote, as will more culturally Conservative ex-NDP and ex-Liberal voters. The problem is I’m not sure the Conservatives will be in a position to win many seats from that. In Northern and Southwestern Ontario the NDP vote is bleeding both ways, nullifying Conservative gains, but also Mark Carney is less toxic than Trudeau. For whatever thinly veiled (or not even veiled) bigoted reasons, Trudeau’s more effeminate nature didn’t always play well. Carney’s a better fit for Windsor and London and the Soo, he just is, and that’s a problem for Poilievre.

The other part of it is that we have significant proof from Leger and Mainstreet, and from the 2024 BC election, that the Conservatives are going to do better with ethnic minority voters this year. But the Conservatives could average 15% swings across the two core Surrey seats, the 6 Bramptons, the 6 Scarboroughs, and the 6 Mississaugas and win legitimately 0 seats. What is far more likely than a sweep of those 20 seats is that the Tories close a lot of gaps, give themselves a more attractive map for 2029, and fail (either mostly or entirely) to win seats under First Past The Post. Call it getting Crombie’d, cause that’s on the table.

The other part of it is that if the Conservatives are doing better than expected in the GTA from what you’d assume they’re doing from Ontario-only numbers, then they’re probably bleeding votes elsewhere in the province. Does that mean Carleton’s in play? That’s what Leger suggested, but I’m more sceptical. But you know who might be in real trouble, if this is true? Jamil Jivani, in Bowmanville-Oshawa North. Purely derived from the province wide numbers he’s up 7% with a 9% NDP vote for the Liberals to squeeze, but there’s also two factors.

Jivani is, famously, JD Vance’s former classmate and friend, at a time when Vance and the Americans are pretty persona non grata up here. Jivani’s also presumably got some of the 3200 direct employees of the Oshawa GM plant in his district, plus a lot of knock on economic consequentials in his part of northern Oshawa. There are probably less than a dozen of seats where the Trump auto tariffs will be more directly consequential than Jivani’s.

There could be a situation for the Tories where they advance in seats they can’t win, they lose seats they won in 2021, and the Liberals in effect run the table of close seats even though the Conservatives increase the number of close seats by a dozen. It would be a moral victory, I guess, but still a loss.

The Liberals survived a major crunch point in the polls this week. Their lead remains stable, and huge. At some point Carney either needs to completely embarrass himself or the Conservatives need to get luckier than a leprechaun shitting rainbows if this election is going to be in any doubt.