Will Justin Trudeau’s Liberals win the 2021 election?

As rumours bound regarding Justin Trudeau dropping the writ for a September election, federal polling firms seem to be pumping out survey results with increasing frequency.

I thought it’d be interesting to run through a few of them. Since Parliament let out at the end of June, I’m going to look at surveys published in July and early August.

		LPC	CPC	NDP	GPC	BQ	PPC
Léger	2 Jul	33	30	19	5	7	4
Angus Reid	15 Jul	33	31	20	3	7	4
Léger	17 Jul	34	29	22	4	7	3
EKOS	18 Jul	35	28	19	5	5	6
Innovative	18 Jul	41	27	17	5	6	3
Ipsos	20 Jul	36	30	20	3	7	2
Angus Reid	25 Jul	33	30	21	4	7	3
Mainstreet	25 Jul	34	32	16	5	5
Abacus	31 Jul	37	25	20	6	8	2
Léger	31 Jul	36	29	20	4	7	3
Angus Reid	1 Aug	35	30	19	3	7	3

In every poll, the Liberals are the clear frontrunner, with the Conservatives breaking the 30% mark 6 out of 11 times and higher than 30% only twice.

Here’s how the average and median look for the 11 polls:

	LPC	CPC	NDP	GPC	BQ	CPC
Average	35.2	29.2	19.4	4.3	6.6	3.3
Median	35.0	30.0	20.0	4.0	7.0	3.0

So roughly 35% for the Liberals, 30% for the Conservatives, and 20% for the NDP, and everyone else with under 10% (or under 5%, as the case may be).

Now, two firms have more than one result in the dataset, so I’m going to combine them so they’re less likely to skew the results.

		LPC	CPC	NDP	GPC	BQ	PPC
EKOS	18 Jul	35.0	28.0	19.0	5.0	5.0	6.0
Innovative	18 Jul	41.0	27.0	17.0	5.0	6.0	3.0
Ipsos	20 Jul	36.0	30.0	20.0	3.0	7.0	2.0
Mainstreet	25 Jul	34.0	32.0	16.0	5.0	5.0
Abacus Data	31 Jul	37.0	25.0	20.0	6.0	8.0	2.0
Léger	31 Jul	34.3	29.3	20.3	4.3	7.0	3.3
Angus Reid	1 Aug	33.7	30.3	20.0	3.3	7.0	3.3

And the adjusted average and median:

	LPC	CPC	NDP	GPC	BQ	PPC
Average	35.9	28.8	18.9	4.5	6.4	3.3
Median	35.0	29.3	20.0	5.0	7.0	3.2

The numbers are still pretty close to the same. The Liberals’ average increased by 0.7 points, but their median remained the same. The Conservatives lost 0.4 points on their average and 0.7 points on their median. The NDP lost 0.5 points on their average but nothing changed on their median. The Greens jumped up a full percentage point on their median score.

The Liberals are still in the lead, ahead of the Conservatives by nearly 6 points and the NDP by 15 points. The Conservatives lead the NDP by a little more than 9 points.

Let’s look at how these numbers compare to the 2019 election:

	2019	2021	Change
LPC	33.2	35.0	1.8 ↑
CPC	34.3	29.3	5.0 ↓
NDP	16.0	20.0	4.0 ↑
GPC	6.6	5.0	1.6 ↓
BQ	7.6	7.0	0.6 ↓
PPC	1.6	3.2	1.6 ↑

We see that the Conservatives have seen the largest drop in support: a full 5 percentage points, compared to the last federal election. The Greens and the Bloc saw losses as well, but not nearly as much as the CPC.

The largest gain went to the NDP, which is up 4 full points since the election. Liberals are up by almost 2 points, and the People’s Party is also up by almost 2 points.

I’m actually not that surprised about any of the gains. This is Jagmeet Singh’s second election as leader of the NDP, so he’s gotten some of the kinks out. Plus, even though his campaign platform isn’t out yet, his rhetoric seems to be more progressive than in 2019, and he’s been trying to capitalize on social media to get his message out.

Justin Trudeau is coming out of the pandemic having led the country through an economic and health crisis. And regardless how much he and his party was directly responsible for, he’ll get the bulk of the credit, which translates into better poll numbers.

Finally, the PPC was a brand new party in 2019, barely a year old. It was inevitable that the party would receive little support in its first election, and increasing support in the last two years shouldn’t shock anyone. But barring any significant scandals with the CPC, I doubt the PPC will gain much more support, either during or after an election this year, if one is actually called.

Now, all that being said, we must keep in mind that we are comparing voter intention with the results of an actual election. There are variables in play that prevent us from saying that these polling results will translate into similar voting results.

For example, let’s take the Angus Reid poll from last week.

LPC	32%
CPC	27%
NDP	17%
GPC	3%
BQ	6%
Other	6%

This is for decided and leaning voters. But notice that if you add them all up, it comes to only 91%. There’s still another 9% of poll respondents who are undecided, didn’t want to say who they’d vote for, or don’t intend to vote at all. Plus, Angus Reid assumes a margin of error of 2.5 percentage points.

That’s a lot of variability.

Not only that, but turnout in the 2019 federal election was 67%. That means that 33% of voters didn’t end up voting. Can you imagine if 33% of those who took the Angus Reid poll decided to not vote after all?

Again, a lot of variability.

Which is why it’s challenging trying to compare polling with election results. But what if we compare pre-election polling with this summer’s polling?

Here are the last 11 polls leading up to the 2019 federal election:

		LPC	CPC	NDP	GPC	BQ	PPC
Nanos	20 Oct	31.7	32.5	20.8	6.0	7.2	1.5
Mainstreet	20 Oct	31.6	32.2	18.3	6.1	7.9	2.8
Research Co.	20 Oct	32.0	31.0	19.0	8.0	7.0	2.0
EKOS	20 Oct	34.2	30.0	18.3	8.1	5.0	3.5
Campaign	20 Oct	31.4	31.3	17.8	9.1	6.6	2.7
Ipsos	19 Oct	31.0	33.0	18.0	6.0	7.0	3.0
Abacus Data	19 Oct	34.0	32.0	16.0	8.0	8.0	2.0
Nanos	19 Oct	31.0	31.5	18.8	9.5	7.0	1.8
Campaign	19 Oct	31.7	31.4	17.4	9.1	6.6	2.7
Forum	18 Oct	31.7	29.9	17.5	8.3	9.0	3.0
Leger	18 Oct	33.0	33.0	18.0	6.0	8.0	2.0

And here they are with Campaign Research and Nanos Research combined.

		LPC	CPC	NDP	GPC	BQ	PPC
Nanos	20 Oct	31.4	32.0	19.8	7.8	7.1	1.7
Mainstreet	20 Oct	31.6	32.2	18.3	6.1	7.9	2.8
Research Co.	20 Oct	32.0	31.0	19.0	8.0	7.0	2.0
EKOS	20 Oct	34.2	30.0	18.3	8.1	5.0	3.5
Campaign	20 Oct	31.6	31.4	17.6	9.1	6.6	2.7
Ipsos	19 Oct	31.0	33.0	18.0	6.0	7.0	3.0
Abacus Data	19 Oct	34.0	32.0	16.0	8.0	8.0	2.0
Forum	18 Oct	31.7	29.9	17.5	8.3	9.0	3.0
Leger	18 Oct	33.0	33.0	18.0	6.0	8.0	2.0

And the average and median for each party leading up to the election:

	LPC	CPC	NDP	GPC	BQ	PPC
Average	32.3	31.6	18.1	7.5	7.3	2.5
Median	31.7	32.0	18.0	8.0	7.1	2.7

Now let’s compare the median scores for each party from the pre-election polls with how they actually did in the election.

	Polling	Election	Change
LPC	31.7	33.2	1.5 ↑
CPC	32.0	34.3	2.3 ↑
NDP	18.0	16.0	2.0 ↓
GPC	8.0	6.6	2.4 ↓
BQ	7.1	7.6	0.5 ↑
PPC	2.7	1.6	1.1 ↓

As far as who came in which place, polling was pretty accurate for the first 3 parties: Conservatives in first place, Liberals in second, and NDP in third. It switched the Bloc and Greens though.

When we look at the actual numbers, however, polling was off on everyone except the Bloc.

Even though both the Liberals and Conservatives ended up with more votes than the polls predicted, the spread between them was much larger. Polling gave them a median spread of only 0.3 points, but the election had the Conservatives winning by 1.1 points, nearly 4 times as much.

Not that it ultimately mattered, however. Because parties win based on seat count and not popular vote, the Liberals were able to use their votes strategically in order to win 36 more seats than the Conservatives.

The NDP and the Greens were both down significantly come election time, 2 points and 2.4 points, respectively. Two things were probably at play here.

First of all, as I stated above, 1 out of 3 registered voters didn’t show up on election day. That can seriously throw a wrench into polling predictions.

And second, with the Conservatives either leading or being in a statistical tie with the Liberals through most of these 11 polls, there’s a possibility that the potential for another Conservative government could’ve spooked some NDP and Green voters to rally behind Trudeau.

So, where does that put us now?

Well, that’s a tough one to predict with preciseness. The race isn’t as close between the Liberals and the NDP at the moment as it was 2 years ago, and it doesn’t seem to be changing. Which means there probably will be less scare voting, and it’s possible the Liberals might even gain some seats.

But let’s do some predicting. Let’s take the median of the most recent polling of the summer and add on the polling–election difference from 2019 and see what that tells us.

	2021	Change	Prediction
LPC	35.0	1.5 ↑	36.5
CPC	29.3	2.3 ↑	31.6
NDP	20.0	2.0 ↓	18.0
GPC	5.0	2.4 ↓	2.6
BQ	7.0	0.5 ↑	7.5
PPC	3.2	1.1 ↓	2.1

So, even if the difference between polling and the election are similar to 2019, the Liberals still come out on top. It’s unlikely that the Conservatives will overtake them, especially when you consider that they got only 31.9% of the vote when Harper lost to Trudeau in 2015, and Andrew Sheer lost to Trudeau in 2019 with 34.3% of the vote.