Is the People's Party hurting the Conservatives?

So it's official (and surprising), the leader of the People's Party of Canada, Maxime Bernier, has been invited to the official leaders' debate. A surprising decision given that the debate commission had initially told Bernier he didn't make the cut. I'm of the opinion that this was the right thing to do as the PPC is running candidates pretty much everywhere and, while it might not elect more than 1 MP (or zero actually), the polls put this party around 2-5%, a score that isn't negligible. It seems to me that should indeed be enough to be invited.

Andrew Scheer's reaction was, to say the least, not positive. See below.

Given that Bernier came super close to winning the leadership of the CPC (he would have if it weren't for the effort of the dairy farmers, let's be honest), and given the general policies of the PPC, the conventional wisdom is that the PPC is mostly hurting the Tories.

I'd tend to mostly agree with this assumption. A poll from Abacus supported this thesis, although it was done right after the creation of the PPC. My own analysis of the data (for instance I looked at the correlation between polls as well as looking at the before/after of polls from the same firms when they decided to include the PPC), confirmed this. Other polls (for instance polls showing how the 2015 votes is distributed currently) also pointed towards the PPC mostly taking votes from the CPC, but not only. I think it's really hard to give an exact estimate. I'd say anything between 40 to 60%. I'd say we can confidently say a majority (or plurality) of the PPC votes are coming from the CPC, but not all.

In my model, I assumed that the PPC was taking 40% of its votes from the CPC, 15% from the Liberals, 15% from the NDP and 30% from new voters. Some might be surprised that I'd even assumed some NDP voters defecting to the PPC but all the numbers analysis I ran showed that there was indeed a link. Remember, chances are that if you are reading this blog, you are a political nerd who knows the full left-right spectrum. For you, it's not logical to jump from NDP to the PPC. But many voters aren't that logical.

Before looking at the current impact of the PPC on the CPC's chances of winning, I need to mention a somewhat counter-intuitive effect of my model. For the sake of illustration, let's round the Conservatives' results at 32% in 2015. If the PPC is polling at 4% and we assume 50% of those votes are coming from the Tories, then they should be at 30% (32% minus half of 4%) if nothing else had changed. So when the polls are currently placing the party of Andrew Scheer at 35%, it means they are up by 3%+2%, not 3%. In other words, they made up for the 2% lost to the PPC by gaining 2% elsewhere. Depending on where this 2% is coming from, it could actually make the Conservative's vote more efficient (imagine for instance the CPC losing the 2% to the PPC in rural ridings but gaining it back in the suburbs). Key word: could.

So what it means is that keeping the CPC constant at 35%, the higher you input the PPC in the model, the bigger the positive swing for the CPC becomes! Again I understand it might seem counter-intuitive but I actually think it's very logical if you think about. Of course the key element here is whether the CPC is indeed really at 35%, but that's another question.

So, is the CPC currently hurt by the PPC? My latest projections have the CPC at 35.2% nationwide and 140 seats, compared to 34.1% and 165 seats for the Liberals (the Liberal vote is just more efficient, thanks to Quebec. Read this piece in French from yesterday if you want to know why the Liberals could make big gains there). Specifically, here what it looks like:

The PPC is at 3.2% and zero seat (Bernier is in a close race in his riding of Beauce). So let's assume that 40% of the PPC votes would vote for Scheer if the PPC didn't exist. 15% would go to the LPC, 15% to the NDP and 30% would simply not vote. Let's assume it's the same in every province (well except in Quebec where we will assume 40%, 10%, 10% and 20% to the Bloc).

Doing so would give the following projections:

So yes the Tories would indeed be higher (they gained 40% of 3.2% but also a little bit more since 30% of the PPC wouldn't vote, thus making the other votes be worth slightly more). But realistically, it's only changing a few seats.

With that said, a few seats could well be the difference between a majority or minority, or a Conservative or Liberal government. So it's too early to categorically state that the PPC isn't affecting this election. But it does seem to be fairly minor right now. Of course, this exercise is full of assumptions. Moreover, with Bernier at the debates, it's possible he'll actually take more votes away from the Tories.

If we instead assume that 60% of the PPC votes is from the CPC and 40% wouldn't vote (so no exchange between PPC, LPC and NDP or Bloc), we get the following:

So depending on the assumptions, it seems the existence of the PPC is costing the CPC between 1 and 4 seats. Alternatively, it means the gap with the Liberals is decreased from 25 seats to 17 in the most aggressive scenario here. So it's definitely not negligible in this case.

In conclusion, I think it's fair to say the PPC could indeed hurt the Conservatives the most. But it's also not currently making the difference between Scheer becoming Prime Minister or not.