Earlier we showed that in addition to today's violent collapse in the reflation trade, another, familiar concern re-emerged in Europe as both the German-France 2Y spread resumed its widening, coupled with the gradual increase in EUR redenomination risk as captured by rising European CDS.
This rising French risk, according to a new report by Deutsche Bank, was driven by the increased uncertainty in the first round (more specifically the risk of Melenchon qualifying for the second round) rather than an increase in risk in the second round (i.e. Le Pen overcoming a 20pp deficits should herself and Macron qualify for the second round). Melenchon’s odds increased together with his polls since late March. At the same time, Macron and Le Pen first round polls declined, while the gap between them in the second round has remained broadly constant.
This sudden rise of the far-left Melenchon, which "threatens" to put him in the second round alongside the much hated by the establishment Marine Le Pen, prompted a concerned Deutsche Bank to ask a whether a Melenchon – Le Pen second round is within the historical margin of polling errors, and is therefore not being captured as a risk in an election where up to a third of potential voters have still not made up their minds how they will vote.
Here is what it found:
Last week, we highlighted that for polls conducted early April, the first round polling error on the gap between two of the three main candidates averaged 6pp historically. We update and expand the analysis in the table below by focusing on the accuracy of the last polls conducted just before the first round for the top four candidates.
This updated analysis does confirm that a polling error of ~6pp on the gap between two of the four major candidates is typical. Currently, the top four candidates are within a 5pp range which makes any second round combination statistically possible. As a result, a Melenchon - Le Pen second round is technically within the range of historical polling errors. So would be a Macron-Fillon second round. However, voters’ commitment to Macron (upper 60s) and Melenchon (mid 60s) appears weaker than for Fillon (upper 70s) and Le Pen (mid 80s). If one assumes that voters’ commitment is indicative of the potential direction of polling error (which is intuitive but undocumented empirically), given current polls, the most likely “surprise” would be a Le Pen – Fillon second round.
And another question concerning Deutsche Bank, and investors: will the Melenchon – Macron gap continue to tighten aggressively?
So far, Melenchon surge in the polls has primarily been at the expense of Hamon. Before the first presidential debate, the combined Melenchon (12%) + Hamon (13%) vote was ~25%, evenly split between the two candidates. Melenchon’s vote has risen to ~19% (+7%), while the Hamon vote has declined to ~8% (-5%) and the combined vote increased to ~27%. This suggests that 70% of Melenchon gains were at the expense of Hamon.
Currently, Hamon polls around 8% of expressed votes, which amounts to ~2.7m votes. About 2m votes were cast in the second round of the Socialist primary in which Hamon beat Valls 60% to 40%. Although it is not possible to link current Hamon voters with the Socialist primary, it is likely that Socialist members who voted in the primary will be amongst the most faithful to the party. This would put an indicative floor of ~6% as to how low Hamon polling could go. In a worst case scenario, one could imagine that only members who voted for him in the primary will ultimately vote for him in the first round of the election. This would put the floor at 3.5%. This leaves another 2-4.5 pp potential votes that Hamon could lose.
The most recent polls give us some indication as to who could benefit from these vote transfer. An Ipsos poll indicates that Hamon voters would have a 2- to-1 second choice preference for Melenchon. In a second round between Melenchon and Macron, Hamon voters would transfer 60/40 in favour of Melenchon. Applying these vote transfers to the 2-4.5 pp potential additional loss of voters would imply scope for an additional 0.5-1.5pp tightening of the Macron-Melenchon gap on the back of further decline in the Hamon vote. This compares to a ~9pp tightening since the first presidential debate on March 20th. Of course, there is still the risk that Melenchon takes votes directly from Macron as well. However, second round voting intentions suggest that he is equally likely to gain votes from Le Pen.
Often, tactical voting becomes more prevalent as the race becomes tighter. In 2012, tactical voting benefited Hollande at the expense of Melenchon. Such shift may not be pronounced this time as polls indicate that Melenchon would defeat both Fillon and Le Pen in the second round. However, it is conceivable that some tactical voting occurs as voters’ commitment to Melenchon (mid 60s) is not high.
Melenchon’s rise in the polls has been one of the key market drivers since the end of March. Further decline in Hamon’s votes is unlikely to support Melenchon to the same extent as it did in the last three weeks. However, the more important point is that the risk around the first round persists as (a) the top 4 candidates are within the historical margin of error and (b) the high level of undecided voters increases the uncertainty of the outcome.
In other words, for those wonder if this is what the second round in the French election may look like, the answer is: it's very likely.