Here we are after 36 days of an election in the middle of a pandemic. Canadians are getting used to those though. So here's the executive summary for you busy people. If you want more, read below.

You have the voting intentions based on the polling average, the seat projections with the 95% confidence intervals as well as the chances of winning the most seats.

The possible outcomes are:

LPC plurality: 60% chances of happening

LPC majority: 18% chances

CPC plurality: 21%

CPC majority: 1% (not impossible! Remember this if it happens lol)


If you want the riding by riding projections, scroll to the bottom of this blog post.


The map



Full version here.


Polls are far from perfect and there is always considerable uncertainty. My simulations use an actual margin of error of over 4%, in line with the effective margins estimated using past elections. Federal elections haven't had a huge polling mistake though, the failures were at the provincial levels. My probabilities have never failed at the riding level, meaning that every candidate projected with 100% have won and every one at 0% have lost. Hopefully the streak continues tomorrow night.


The most likely outcome is a Liberal minority. If polls were to be wrong, a CPC plurality is slightly more likely than a Liberal majority, but both outcomes wouldn't shock me tomorrow night.



1. The polls


This election was marked with a quick rise of the Conservatives. After about a year and a half of trailing the Liberals significantly (since covid-19 stared really), including during this summer, the Tories just caught back to the Liberals as soon as the campaign started. It looks like the unhappy Conservative voters (in Alberta for instance) just decided to rally behind O'Toole, at least for a while (before some decided to leave for the PPC).



We even saw the Conservatives take the lead at some point, although the average was heavily influenced by the very volatile IVR polls (some of which showed the CPC with a 8 points lead at some point, lol). Online polls (like Leger or Abacus) have been stable for weeks. IVR polls also disagree with online one for the age crosstabs. IVR polls have the Tories doing well among the 18-34 but trailing the Liberals in the 55+ while online polls have the Conservatives slightly ahead with this important demo (they vote more). The two effects cancel out for the overall numbers but it's still weird. Using the average, here are the voting intentions by age:



If the turnout is low tomorrow and skews heavily towards the 55+, it could explain some potential errors we'll observe. It should be noted, however, that the Liberals are making gains with the 55+ this election. They have a fairly different coalition of voters compared to 2015 when they dominated the 18-34. Speaking of demographics, the Liberals are also losing a little bit among the university educated (NDP gains) and among men (they gain among women which creates a really huge gender gap).

There is some momentum for the Liberals at the end although, there as well, it's mostly driven by the IVR polls. Ekos has the Liberals just skyrocketing in Ontario in the last few days. Are they actually picking up a trend other polls haven't? Hard to say, Ekos is notorious for having weird (wrong...) regionals at times. Also, even if the Liberals indeed have the momentum, the high number of ballots already cast could ultimately mean the overall results won't be as favorable to the Grits as the last minute polls (see below). But it does look like Trudeau managed to increase in the last days in Ontario, we saw this trend in many polls. See by yourself in this graph of Ontario.


It is notable that the CPC never took the lead there, which explains why I never had them as overall favorite. It did come close at one point though.

Quebec was quite different with a stable, albeit slowly declining Liberals. The Tories had the momentum for a while but the debates (and the controversy) ultimately helped the Bloc.



Finally, in BC, this is by far the worst province for Trudeau. His party has been declining steadily for the entire campaign.



We obviously need to talk about the People's Party of Maxime Bernier as that was one of the events of this campaign. They started low, so much so actually that they didn't qualify for the debate even though they only needed to poll at 4% in average (note: they were over 4% in many polls but not in average). They are now around 7-9% in some polls, it's insane. They could finish second in Alberta for instance but they also poll well in Atlantic Canada or Ontario. We also know the PPC voters are younger, more male and less educated (not necessarily high school diploma only but definitely fewer university graduates). Bernier has definitely managed to gather the 'angry mob'. He's taking voters from the CPC but not only. Multiple polls have shown that many 2019 Green voters were now PPC. I also suspect the PPC is just turnout out new voters, people who never had a party to vote for. This is fascinating. He could well play spoiler in a few ridings.


1.2 Advance turnout and mail-in ballots.

We know a large number of people have already cast their ballots, either during the advanced polls (5.8m) or by mail (900k ballots received as of today, but 1.2m kits sent). Given that most pandemic elections have seen a decrease in turnout, and given that people more and more vote in advance, I believe this 6.8-7m votes could represent as much as 40% of the total ballots, maybe more. Extrapolating from the % of people who declared having voted already in multiple polls, I believe this estimate of 40% is realistic. That means we can expect a total number of votes between 16 and 17.5m if I had to estimate. This represents a drop of 1 to 2.5m ballots compared to 2019 and would likely mean a turnout of around 60%, down from 67% in 2019 and 68% in 2015. If there is a surprising outcome tomorrow night, I'm willing to bet it'd be caused by a low turnout.


What do polls tell us about the people who have already voted? Using data from Angus-Reid, we see the Conservatives are likely to be ahead after the advance voting and mail-in ballots (even though the latter favor the NDP and Lib more). I have myself tried to reverse engineered the numbers among the 'already voted' crowd and we see good numbers for the Conservatives in BC but less favorable ones in Ontario. Given the last minute trend in this province, O'Toole should be quite worried. If his party is already trailing by 6-7 points before the new votes on Monday, things could get ugly.


Other pollsters have provided us with some info. Léger showed us that CPC voters were more likely to already have voted in advance or would then vote in-person while NDP voters were the most likely to vote by mail. However, their most recent poll showed that the Tories were the party with the lowest percentage of 'voted at the advance polls'. Ipsos confirming the NDPers favoring mail-in ballots more while also showing the Conservatives not doing that great among the already voted crowd. So definitely some conflicting info here. Moreover, Earnscliffe was showing both the CPC and LPC at 27% among the 'already voted' crowd but the Tories with a big lead in Ontario (34% to 26%). Finally, Abacus is showing a decent lead for the CPC among the people who had already voted (my calculations here).


So, all in all, I'd say that advance polls likely favor the Conservatives a bit. Given that, as we mentioned, about 30 to 40% of the ballots have already been cast, it could well protect the Conservatives from a collapse (in the eventuality that Ekos is right). Even if they trail by a lot in the votes cast on Monday, the advance polls should allow them to remain above 100 seats at least. Will it be enough to prevent a Liberal majority? It's less certain. On the other hand, if same-day turnout tomorrow is anemic, a Tory plurality is a lot more possible.


2. Seat projections

First, let's be very clear about one thing: seat projections are very sensitive to even small polling error. While people commonly accept polls to be off by 2-3% (within the margins of error), such deviation can cause the projections to be off by 30 seats. Mapping voting intentions into seats is tricky. I'm mentioning it because while the overall numbers look good for the Liberals, people should keep in mind that it only takes a smallish error in Ontario (let's say 3 points off) to change the party finishing first.


I'd like to mention, again, that my projections never had the Tories as favorite. It got tight at some point but the Liberals were always favored. I do not understand projections websites who had the CPC with over 50% chances of winning at any point. To me this is just the result of assigning too high of a weight to a few volatile IVR polls.


What we see is that Atlantic Canada is likely to remain mostly red. As a matter of fact, it could be even redder than my projections here if the Mainstreet riding polls are right. Quebec could have been the source of many gains for Trudeau but the campaign (and the English debate's controversy) changed that. At this point, few seats should flip in la Belle Province. With that being said, we have seen some evidence (riding polls, regional numbers, demographic numbers) indicating that the Bloc vote is more urban than last time. So they could gain seats in the South shore of Montreal while losing Jonquière to the CPC. See the map of Quebec below.


It very much looks like 2019. it gets more interesting if we zoom in.


The Bloc regain Hochelaga and Sherbrooke. They could also retake the two missing South shore ridings. The NDP is dreaming of Laurier-Sainte-Marie, which is possible assuming the Bloc doesn't split the anti-LPC vote. Notice Trois-Rivières going CPC which would be a big deal. Speaking of which, this party has legitimate chances in Beauport-Limoilou and Jonquière but will need to make sure not to lose Beauce to Bernier.


Ontario remains the main battleground. The Tories could well decrease their vote share for the 3rd election in a row. They are now projected to receive a much smaller share of votes than the sum of the previous PC and Alliance. This is hugely problematic for this party and especially for Erin O'Toole who is from the GTA and based his entire campaign on making gains there. Quite frankly, if the CPC doesn't make gains in Ontario, I can't see O'Toole surviving as leader. The base is not happy with him and his 'centrist' approach (he even has a carbon tax!!), they only tolerated it in the hope of making gains in the GTA. And so far, we have no reason to believe it'll be the case. The provincial numbers aren't good enough and getting worse in the last minute. The sub-provincial numbers (some polls have data for the GTA only) have never really indicated major gains for the Tories except one Leger poll in early September. The riding polls in the region have also not been good. The Tories will also have to deal with vote splitting with the PPC, especially in the Southwest of the province. The only good news for O'Toole is that unless he completely collapses, the Liberals are unlikely to make the gains they need for a majority -- they will just keep the seats they already have.


Some have argued the rise of the PPC would make the CPC's vote more efficient (at the same %). I can see the logic behind it. But again, the sub-provincial numbers haven't shown any evidence the Tories are about to make gains in the GTA. Maybe the only indication is the seat clusters analysis from Innovative showing the CPC closing the gap with the Liberals in those competitive Ontario seats.


As you can see, it's still mostly all red in Toronto and the suburbs. But you also see a lot of light red meaning the races are close. The Tories are really within 5 points of flipping 10 seats.


In the north, this is an interesting 3-way race. The NDP has good chances of flipping Nickel Belt and Thunder-Bay but it's also possible the NDP under Singh is just too urban, to woke and too young for that. We'll see.



The Prairies and Alberta shouldn't be super interesting. A few races here and there, but that's it. Although there are some indications that the NDP could be making some gains. Coupled with the possible rise of the PPC, I wouldn't be shocked to see the NDP making a few unexpected gains. As for the Liberals, they will try to win seats in Alberta this time around and odds are in their favour.


BC was supposed to be the source of the majority just 2 months ago. But that's the one province where the Liberals have dropped almost continuously during the campaign, including at the end. It's close to a 2011 collapse really. And the riding polls have also been mostly terrible. At this point, I expect the Liberals to keep their few urban seats in Vancouver and the suburbs but not much more. It's interesting how different the trends have been between Ontario and BC.


Yes this is orange (well yellow on my map) Vancouver Granville. Might be surprising but there are good reasons to believe the NDP will take this riding after JWR retired. Coquitlam-Port Coquitlam is literally tied at 33.8% between the CPC and Liberals and you need to look one decimal further to have the LPC winning it. still, very few safe seats for the Liberals.


So, overall, the Liberals are projected with a minority. Given that polls are far from perfect, we can't exclude a majority or a CPC plurality (whether that would be enough to govern is another question). The chances are 21% and 18%respectively. It might seem surprising that a majority isn't seen as more likely but this is partially the result of the crazy LPC vote efficiency. They win many ridings by small margins (or they did in 2019) but that also means their next potential targets are further away. With Quebec being so close between the Bloc and the Liberals, it's hard to see where the majority would come from. As I was explaining here, it would require a pretty significant polling error. An error in line with the one in Quebec during the 2015 election when the Liberals won their surprising majority. I really believe that the Quebec and BC numbers make a majority out of reach this time. My model already allocates slightly more undecided to the Liberals (incumbents) so I doubt I'm underestimating this party, although we never know after 2015 and 2019.


If we were to witness a Liberal majority, it would likely start with Atlantic Canada being very red, like 28+ seats for the Liberals (and likely zero for the NDP). Then Quebec with the Bloc failing to make gains in Hochelaga, Sherbrooke, etc. As a matter of fact, we would see a few seats in the 450 (Montreal's suburbs) switching from Bloc to Liberals. Then, in Ontario, we would see the Liberals well over 40%, winning everything in the 416 and most of the 905. Add a few gains thanks to vote splitting due to the PPC. The rest of the night would need to wait for a few urban pickups in Alberta and a few gains in BC. There is a path to 170, it's just not the most likely one based on all the data we have.


A CPC plurality would have to involve surprising gains in the Atlantic, similar to the PC victory last month. I'm talking of at least 8 seats there. In Quebec, the Tories would win Beauport-Limoilou and Jonquière while keeping everything else. They could be in the race in one of the Abitibi ridings. In this scenario, they need the Bloc to be strong around Montreal and take seats away from the Liberals. Then, the big moment, the GTA. Instead of being red, we see the Tories winning a lot in the 905. King-Vaughan, Aurora-Oak Ridge, etc. Let's be clear here, that would mean the polls were quite off in average. Instead of Liberals +7, it's Liberals +2 or even tied! From there, the West remains heavily CPC (similar underestimation as 2019; PPC voters deciding, in the last minute, to vote CPC) and all urban hopes in Alberta are mostly crushed. BC falls mostly in line with the projections but the CPC wins over 20 seats. In this scenario, the same-day turnout is likely quite low.


If we think of outcomes as who has the confidence of the house, then we likely need to look at whether LPC+NPD>169 or if CPC+Bloc give a majority. The simulations show that the Liberals should have the confidence of the house, either alone or with the help of the NDP, 75% of the time.


Let's talk PPC. I don't have this party projected as winning a single seat. But models like mine are most likely not the best when dealing with a mostly new party experiencing a surge. Using my model, polling data and other data (vaccination rates, searches on Google), I believe we could see a PPC MP in Elgin–Middlesex–London where Chelsea Hillier is running, the daughter of a provincial MLA who was kicked out of the PC caucus for being anti lockdown. She has the name recognition and the searches on Google are through the roof. Mainstreet did poll that riding and only found us a distant 4th though. Still, one of the PPC's best chances. Otherwise, look in rural Alberta for ridings like Grande Prairie-Mackenzie, Lakeland, Medicine Hat-Cardston-Warner, Red Deer or Battle River-Crowfoot. I honestly believe the PPC might indeed win a seat tomorrow but I don't want to subjectively modify my model.


2.1 Other projections

You can find other seat projections yourself, there are many. But I wanted to mention ASI Polly who uses social media buzz (they don't provide a lot of details). They have shown a Liberal surge towards the end, thus mostly confirming what the IVR polls have seen.

Databright also seems to look at social media and is finding a situation very similar to the polls.

Also, if we use Google Trends, we see Trudeau dominating but not enjoying a particularly strong end of the campaign. If we instead look at the parties, the PPC is doing incredibly well.


In conclusion, there is a massive consensus for a Liberal minority. With that being said, I can't remember the last federal election where the results fell in line with the consensus! So expect the unexpected tomorrow!


Detailed projections


Final Projections Canada 2021 by bryanbreguet on Scribd

This will likely be my last update before the final projections on Sunday. The race remains tight but the Liberals have a clear advantage thanks to their vote efficiency. With that being said, the chances of a LPC majority are half that of the chances of a CPC plurality. Both 'surprises' would only require a minor polling upset but the Tories do seem closer to finishing first than Trudeau is from a majority. Vote efficiency works both ways: it makes you win a ton of seats by small margins but it also means the next list of winnable ridings is further away.


So here are the projections and I look at what numbers we'd need in Quebec and Ontario to see a Liberal majority on Monday (or Tuesday I guess since mail-in ballots won't be counted until the next day).


The wild card is the PPC who is clearly enjoying some momentum -- and the events in Alberta will help. Can the PPC hurt the Tories enough to split the vote? Or maybe straight up stealing seats in Alberta or Southwestern Ontario? Not impossible. But I admit that a model such as mine will almost surely miss this. However, looking at other sources (Google trends, etc), I personally think the odds for one PPC seat are not bad. Ekos has also indicated that those voters are motivated and committed, so it's unlikely they'll switch to the CPC on Monday.


Map


Full version here.



Detailed projections

Proj Canada 17 September 2021 by bryanbreguet on Scribd


What would it take for a Liberal majority?

I had looked at this previously, back when the Liberals still had a pretty massive lead over the Bloc in Quebec. Since then, the lead has mostly evaporated and the lead over the CPC in Ontario has also shrunk a little bit. This means a majority is far further for the Liberals. See the new probability curves.


You might be wondering why the chances aren't higher at +4-5 (which is where the Liberals currently stand in Quebec and Ontario) since my projections above show a 15% chance of a majority. The reason being that the simulations in the projections use +4-5 as a starting point and then apply the effective margins of error. That means the 15% chances mostly come from the simulations where the Liberals outperform the polls. But if they really do win Ontario by 5 points and Quebec by 4%? The chances of a majority are almost zero.


We also see the limited usefulness of Ontario for the Liberals. Outperforming the polls there wouldn't help Trudeau much as he is already winning most of the seats. He would just prevent some losses. Quebec, on the other hand, is more promising. If the Liberals could win Quebec by a comfortable margins (10 points or better), then a majority is all of a sudden a lot more likely. And just to be clear, the orange line does indeed show that even winning Ontario by 18 points over the CPC wouldn't guarantee a majority. But you need to interpret this as the Liberals staying the same everywhere else (including Quebec) but winning Ontario like 45% vs 27%. Again, the Liberals already win a ton of seats in Ontario, at some point there are diminishing returns.


What this shows is the Liberals need a systematic polling error. They can't hope to simply outperform the polls in one big provinces. They have fallen too far in the West, especially in BC. Looking at the simulations, the Liberals need (in average), at least 35.7% Canada-wide for a majority. That means a polling error of about 3 points. That is more than the average error observed during federal elections (note: I'm not talking of a 3 points deviation from any given poll but from the polling average). Given that a good share of the ballots have already been cast (and we have no reason to believe the Liberals did better than the Tories during the advance voting), the Liberals need a clear national lead in the polls by Sunday. Or they'll need some crazy vote efficiency but we should be reaching a point where the Liberals can't really improve on their efficiency from 2015 and especially 2019.

A ton of polls over the last two days but no big changes to the projections. We are so close to 2019m in the big provinces (Quebec, Ontario) that there could be that many seats flipping.


Where is the uncertainty? Well Atlantic Canada seems to be breaking for the Liberals even more than before, which would be in line with the riding polls of Mainstreet -- Literally every single poll done there has had the Liberals ahead! Quebec still sees the Bloc climbing back onto the Liberals. With a vote that appears to be more urban than in 2019, that looks good for Blanchet to win the South Shore and Hochelaga.


Ontario is looking more and more like 2019 although the big increase of the PPC might add some canceling swings. For O'Toole, he has to make gains there or his entire strategy of putting the party at the center will not be tolerated by the base.


Finally BC is the one bright spot for the Conservatives, especially the riding polls. Liberals do appears to be in trouble there although you have enough variations between polls to make it officially a 3-way race.


Anyway, here are the projections.



It should likely look like this until the actual election. Which means the Liberals are heavily favored to win a minority. The next most likely outcome (so most likely 'surprise') would be a Tory minority and only after that a Liberal majority.


Map


Full version here.


Detailed projections.

Proj Canada 15 September 2021 by bryanbreguet on Scribd

The post-debate surge for the Bloc appears to be real. All of a sudden, Yves-François Blanchet is back over 30 seats and can even dream of 40. While Trudeau and the Liberals are gaining in the Rest of Canada (a little bit or a lot, depending on whether you ask Abacus/Angus-Reid or Mainstreet), they are dropping too fast in Quebec to secure a majority. The good news for them is that a CPC plurality is now less likely -- not impossible but definitely harder to imagine.


Notice the 95% confidence interval for the Bloc, now starting at 16 (it used to be in the single digits) and going as high as 44. Urban ridings like Hochelaga are now back with the Bloc.


The chances of a majority are back down around 15%. Some polls (Mainstreet) do have the Liberals close to it but others (and the average) show this party quite far from the magic number of 170. Right now, the Liberals would be losing seats in Atlantic and Ontario. The gains they could make in the Prairies and BC don't compensate. And with Quebec not 'cooperating' anymore, it's really hard to see where the remaining 19 seats could come from. Short of a polling error of course.


Map


Full page version here.


Detailed projections

Proj Canada 13 September 2021 by bryanbreguet on Scribd

A few polls added, some good for the Tories (new Nanos with the lead in Ontario), some very bad (Mainstreet). Overall a stable situation even if the Liberals make a few gains. We are still in the same situation where 2 points in Quebec and/or Ontario could change everything. So I really don't have much more to say than the Liberals would be favorite tomorrow but the chances of a CPC plurality are far from zero and about twice the chances of a LPC majority.



Map


Full page version.


Detailed projections

Proj Canada 10 September 2021 by bryanbreguet on Scribd

Who won the English leaders' debate?
Justin Trudeau
Erin O'Toole
Jagmeet Singh
Annamie Paul
Yves-Francois Blanchet
None


Which party will you vote for?
Liberal party
Conservative party
NDP
Green party
Bloc
Other/Ind.

 Vous savez comment ça marche, deux sondages. Merci de participer.


Qui a remporté le débat des chefs en français?
Justin Trudeau
Erin O'Toole
Jagmeet Singh
Annamie Paul
Yves-François Blanchet
Aucun
Pour quel parti allez-vous voter?
Parti Libéral
Parti Conservateur
NPD
Parti Vert
Bloc Québécois
Autre parti/indépendant