Yesterday was a good day for the polls in France and, by extension, for my projections and myself. The polls for the presidential election in France were spot on and therefore so was I. To be fair, it's much easier to simply aggregate the polls than to convert them into seats anyway.
Let's not forget the BC election though. We haven't had many polls (yet?) but I made some substantial updates to the model. First of all, I took into account where the Green and especially the BC Conservatives were running candidates. The Green have 83 candidates (out of 87) while the BC Cons only have 10. Honestly I should just drop the Conservative from my projections but they were already included, so it's actually less work to leave them there.
I also made some adjustments to some ridings. Some adjustments are coming from the lack of a Conservative candidate (in ridings where this party got over 10% last time around), others are coming because I have reasons to believe my projections were wrong. One such example is in Victoria Beacon-Hill where I previously had it leaning Green. But that was with the built-in bonus of having the Green leader running there. It's not the case anymore and my estimations show Jane Sterk did benefit from a significant bonus. On the other hand, I now have the Green ahead in Cowichan Valley as it appears fairly obvious the Green are running a better campaign than the BC NDP (whose base was kinda split during the nomination process, with the former campaign manager now running as an independent!).
You can find the updated projections on the BC Election 2017 page and you can use the updated model in the simulator.
I wanted to try a little exercise related to the French election. France uses a run-off system where, if no candidate gets more than 50% of the vote in the first round, the top 2 then go to a run-off 2 weeks later. I wanted to simulate the effects on the BC election.
To do so, I used my projections as the baseline results and I redistributed the votes of the other parties in every riding where no candidate was getting a majority. I used the second choices provided by Mainstreet and Ipsos.
The results? See below:
The BC NDP would go from having a small lead to being pretty much guaranteed a majority. For the Green it wouldn't change anything. It's because they are winning their three ridings "comfortably" and they are the main second choice of the main two parties.
Only 39 ridings are currently projected to have a a candidate over 50%. As for being in the top 2 (i.e the run-off), the Liberals qualify 79 times, the BC NDP 85 and the Green 10. These numbers might seem crazy but remember there are only really 3 parties, so it really limits the possibilities.
An example of a riding that would flip from Liberals to NDP with the French system is Boundary-Similkameen. There are 5 ridings affected in total: The one mentioned plus Courtenay-Comox, Maple-Ridge Mission, Surrey-Panorama and Vancouver Langara.
I think that beyond the fun exercise, this blog post should illustrate how close this election is. You flip 5 ridings and you have a NDP majority. If it remains really like this until May 9th, my final projections will be quite uncertain.
