dimanche 5 mars 2017

Qc125 electoral projection for March 5th, 2017

Facing a steep rise of readers, this blog will occasionally be published in English as well as in French. Welcome to my new English readers!

This blog post contains the results of the latest Qc125 projection, first published in French on March 5th 2017.

Quebec polling firms have not released voting intention polls for the month of February, but I felt a new projection was in order considering the new and reshaped electoral map had officially been approved by the Commission de représentation électorale (CRE).

Four LPQ ridings (three solid LPQ ridings and one LPQ swing riding) have been merged into two:

  • The Montreal liberal strongholds of Outremont and Mont-Royal are merged into Mont-Royal-Outremont, effectively removing one of the safest LPQ riding in Quebec;
  • The Laviolette (solid liberal) and Saint-Maurice (swing liberal) ridings north of Trois-Rivières are merged into Laviolette-Saint-Maurice.

The CRE keeps the seat total of the National Assembly at 125, so two newly added ridings were also announced last week. Both are located north of Montreal:

  • Les Plaines, which neighbours Blainville and Terrebonne (both fast growing suburbs of Montreal);
  • Prévost, just north of Saint-Jérôme, the largest city of the Laurentian region.

The new ridings will most likely be a battle between the Parti québécois and the Coalition Avenir Québec. The LPQ will not be competitive in either these new ridings.

Without further ado, here are the results of Qc125's electoral projection for March 5th 2017:

The Liberal Party of Québec is still favoured to win the election had one been held this month, but the Liberals' lead is tenuous at nest. Premier Couillard's party wins barely more than half of the ten thousand simulations - 55.6% to be exact. The Liberals have a 2.7% probability to win a majority government and 52.9% to win a minority.

Close behind the Liberals is the PQ with 42% of simulations going its way. Strangely enough, even if the PQ wins fewer simulations than the Liberals, its probability to win a majority is twice as high as the Liberals': 5.1%.

Yes, that is indeed the case: one out of twenty simulations is a PQ majority, with barely 30% of the popular vote.

Lagging behind is François Legault's CAQ with 0,4% of simulations going its way. All CAQ victories are minorities.

There is also a 2,0% probability of a deadlock - a tie for the seat lead. Most of these scenarios are LPQ-PQ ties, but other, less probable, scenarios are also possible, such as a LPQ-CAQ tie and three way LPQ-PQ-CAQ deadlocks. (And wouldn't that be interesting.)

Here are the seat and popular vote averages for the projection's 10k simulations:

The LPQ wins an average of 49.6 seats with 32,8% of the popular vote. Not far behind is the PQ with an average of 46.9 seats and 28,3% of the vote. As one may see here, a close three way race combined with the new electoral map results in a significant seat projection tightening.

Just a few points behind the PQ, the CAQ wins an average of 25.2 seats with 24.8% of the vote. We will see in future post how CAQ hopefuls should keep the faith despite this seat discrepancy: at 24.8%, the CAQ is just a few points below an important threshold: a 2-3 point swing with the PQ would completely turn this race upside down. In fact, the CAQ seat total is extremely volatile: it can go down to an abysmal 10 and as high as 45. In such three way races, our FPTP system generously rewards parties that go above the 25% mark.

Finally, Québec solidaire wins an average of 3.3 seats with 9.4% of the popular vote. QS leads in the three seats it currently holds and has potential in other seats surrounding its Montreal base.

Here is the popular vote probability function:

According to this graph, the Liberals' popular vote can go as high as 36.5% and as low as 29%, although both extremes are much less probable than the centre of this almost-gaussian curve.

Here is the seat distribution as a function of the popular vote:

I will have more in depth analysis of the projection's details in the coming days, but one can see straight from this graph that the PQ vote is the most effective of the three major parties (the dark blue dots shape the steeper slope). This is a double edge sword for the PQ: a few points above its current average and they become the hands on favourite to win the election, but few point lower its average and it risks becoming the second official opposition behind the CAQ.

Here is the seat probability distribution:

It is curious indeed that the Liberal's curve is a camel, the CAQ's curve is an elephant (steep on the left, gentle on the right) and the PQ's curve is almost gaussian.

This means the following:

  • The Liberals have a higher probability of going above its mean seat total or fall below it. Either they win big or they lose big.
  • The CAQ should easily win at least twenty seats and its probability to win more decreases gently.

The regional seat distribution is hardly surprising: the Liberals lead Montreal, Laval and Gatineau. They barely hold on to half of their seats in the Quebec City region.

The PQ leads in the surrounding suburbs of Montreal as well as the deeper regions of Quebec.

The CAQ leads in Quebec City, the Drummondville (Centre du Québec) region and holds its ground in the 450-suburbs. Its potential to gain seats is mostly against the PQ in the Montreal suburbs.

Finally, QS leads its ridings surrounding its Plateau-Mont-Royal base.

For my new English readers: I will not translate every single Qc125 blog post as it is very time consuming and I already have a full time career. However, I will try to translate as many of them as I can, and I will for sure translate new projection posts in the future.

Thank you for reading Qc125. Connect with me on Twitter and/or Facebook by clicking the links below.

Have a great day!

* * * * * * *

Qc125 sur Twitter: @Qc_125.

Qc125 sur Facebook: @Qc125

Qc125 sur Google+