As long term followers of DataSwarm may know, we have used the system for geopolitical prediction work, including predicting elections. We correctly predicted the result of the previous French Election (Macron won), but as we discussed at the time, we thought Le Pen would do better. But the reason she did not is interesting for predicting this election as well.
To recap – the first round of the French election is to find the two candidates who get the most votes overall, and they go forward into the second round. The voting in the second round then chooses who forms the next Government. The first round is shaping up much the same as the first round in 2017, i.e. Macron vs Le Pen, with Melenchon in 3rd place (see chart above). And despite all the same old dramas of who is up on any particular day, the result of the 1st round has been clear for some weeeks – barring some major unknown factor, Macron and Le Pen will be the candidates moving forward into the second round, and they will be within a few % of each other.
(Update Monday 11 April – Macron 27.6%, Le Pen 23.4% – 4.2% points difference, 51% of vote between them)
And that is where it gets interesting, and prediction becomes hard. Because what’s interesting is not the c 50% of the public vote that Macron and Le Pen get in round one, but the other c 50% who did not vote for them in Round 1. Where do those people go now?
The votes for all the other candidates in Round 1 now splits between Macron and Le Pen – and predicting how that all splits out is non trivial. There are only 2 weeks from Round 1 to Round 2, and Easter falls in the middle, and there is a major candidate debate during in this fortnight, in which last time Le Pen came an (unexpected) cropper. That all makes it hard to track shifting voter sentiment compared to the 1st round, where you see the vote for the candidates tracking over a longer time.
At the risk of more hostages to fortune, this time it is looking far closer at present – and given the current global geopolitical situation and the difference in Macron’s and Le Pen’s policies, this election has potentially far more significant impact than 2017.