March 2021 Projections Update

There were two polls in March, one from Ireland Thinks, and one from Red C. Unfortunately, despite doing cool things like decimal scores, the Ireland Thinks poll once again did not publish a regional breakdown, so the seat changes here are purely based on Red C. They might be more interesting otherwise (Ireland Thinks had FG, GP, IND notably lower, and FF, SDs, PBP notably higher than Red C did), but we can only work with what we’ve got.

Regardless, Sinn Féin moved ahead of Fine Gael on the national RPA (which does factor in Ireland Thinks) for the first time in a long time – they are now at 29% to FG’s 28.5%. This gap has been slowly but steadily closing since last summer, but in recent months this is more indicative of a drop for FG than a rise for SF. To illustrate – on the 9th of September 2020, when the gap was biggest, FG’s RPA was 33.8%, and SF’s was 28.1%. Since then, SF have risen about a percentage point; FG have dropped over 5%. The below graph illustrates this movement since GE 2020:

I’m going to digress here to talk a bit about RPA. This is primarily for the benefit of the gigantic doofuses at Aontú, who not only lifted my work without credit or attribution (always a sign of a highly professional political outfit) but also have clearly demonstrated to everyone that they either don’t understand what the word “consistently” means, or how RPAs work (or both). To demonstrate the falsehood of their claim, here’s the RPA for Aontú measured against PBP:

Firstly, we’re talking fractions of a percentage in polls that have an MoE of usually ~3%; taking that as concrete is seriously not knowing how to read polling. All the RPA says is that on a weighted average of the last 10 polls, Aontú are ahead by a number (0.085%) that is so far within the margin of error that it may as well not exist. Secondly, neither are “consistently” ahead in the RPA – Aontú and PBP have been swapping places from miniscule swings over the last five polls. Finally, if your polling average is less than 2.5%, you probably have slightly bigger things to focus on than the performance of other parties. Anyway. Back to the main stuff.

Despite recent polling not looking good for the government, Fine Gael and Fianna Fáil see no overall change this month. Nor do the Green Party, but they don’t have much further they can fall, seat-wise at least. A lot of constituencies are moving closer, as I’ll discuss below, but the only net change from February is a Sinn Féin loss. Granted, that’s in an extremely competitive constituency that I have flagged before as one that will be swinging back-and-forth a lot over the coming months. The overall map is below:

There are three constituencies where the model projects changes from February:

• Cork South-West (SD +1, SF – 1)
• Kerry (FF +1, FG -1)
• Mayo (FG +1, FF -1)

I flagged last month that this constituency was going to bounce around, so this isn’t a great surprise. The transfer friendliness of Holly Cairns (SD) is the difference maker here. While FPV indicates that she will poll a close 5th, the transfer analysis has her overtaking SF and FF. This would result in an SF elimination, comfortably seeing her home. If Cairns isn’t able to overtake the SF candidate, her transfers would put them over the top.

On the other hand, FF’s Christopher O’Sullivan remains ahead of SF and the SDs on FPV (and ahead of SF even with transfers), but this shouldn’t matter – to keep his seat, O’Sullivan needs to beat FG; his performance against SF or the SDs is irrelevant due to how the model projects transfers will break.

With Michael Collins (IND) still looking very comfortable, it seems that behind him there will be one seat for the left and one seat for the right. The overall point here is that this should be viewed as two individual battles – SD vs SF, and FG vs FF. Whoever wins each of those brackets should take a seat.

This was another one flagged last month as a potential mover, and a small downward shift in the FG vote has changed the final seat to FF, although this is by a very small margin. The FF vote has dropped, but the FG vote has dropped more and is now just a touch behind the threshold where they can expect to comfortably bring in two candidates. This remains far from definite; an improved geographic division of the constituency could well bring home a second FG candidate, but for now, FF have the slightest of edges. This, of course, assumes FF only run Norma Foley – given the weak FF to FF transfer rate in Kerry, a running mate may be a mistake, though whether or not Foley can pull in first preferences in the south and west of the county (something the model, by necessity, assumes she can do) remains to be seen.

The Independent numbers are rising as well, it’s worth noting, so Danny Healy-Rae is very much in the running. Given that the Healy-Raes have been masterful at carving up the vote in Kerry in the past, we are very much approaching the area where Michael might have enough of a cushion to reduce his own vote to give Danny a boost. If this keeps up, the final seat is going to be very, very close.

I do have questions overall about the frequency with which the model is predicting three seats for a single party in certain constituencies. It shows this in Donegal, Dublin South-Central, Louth and Waterford for SF, and in Dublin Bay South, Dún Laoghaire and now Mayo for FG, and I want to do some rationalisation on how the model optimises candidate numbers for April’s update, which should help eliminate any three-seat wins that are noise.

But with that said, polling implies FG will take over 50% of FPV in Mayo, which in a four seater is enough to put three TDs in, depending on how things break with other parties. FG taking an overall majority of votes in Mayo is well precedented – they broke 50% in the constituency in 2007, 2011 and 2016. They haven’t won three seats here since the constituency moved from 5 to 4 seats, but the key thing is the projected collapse in the FF vote – Dara Calleary’s seat is at serious risk. With that said, similar to Kerry, this is very close and there’s a lot of ways FG could fail to capitalise. As mentioned above with Foley, a running mate is a risk, but without one, Calleary would need to be able to collect much more FPV in the south of the county than he did last time out.

Other Notes

• Cork East: Although Labour aren’t really improving anywhere at the moment, the continual cratering of FF support is opening potential doors. In Cork East, if FF were to run two candidates, Labour’s Seán Sherlock will have a very good opportunity to keep his seat – but this does rely on FF making what the model would view as an error.
• Dublin Bay North: One of the messier constituencies has become a little clearer this month, with the SDs Cian O’Callaghan moving into a more secure position. Himself and Labour’s Aodhán Ó Ríordáin seem to have a bit of fresh air between themselves and their rivals for the final two seats.
• Dublin Bay South: Fine Gael moving more comfortably into three-seat territory here – SF don’t have a chance at a second candidate, and everybody else who could realistically challenge them – the Greens, FF and Labour – seem to be going backwards.
• Dublin Fingal: This one is just a mess right now. Joe O’Brien (GP) is still holding on, but Duncan Smith (LAB) is closer than before, although the numbers for both are declining, which may ultimately open things up for a second FG candidate. There’s also a hefty independent vote from 2020, just under 13%, that split between a large number of candidates. It’s very difficult to account for where this will go in future.
• Dublin Mid-West: Yes, the model still shows Paul Gogarty winning a seat.
• Dublin North-West: This is getting a bit more interesting – the gap between a second SF candidate and the SD’s Róisín Shortall has shrunk substantially. While the SDs aren’t surging in Dublin – their support is more or less at GE 2020 levels – if positive trends continue this projection could move back in her favour shortly.
• Galway West: Galway West continues to show weird results in the model; now it’s indicating that Éamon Ó Cuív could be in trouble with FF floundering. I personally don’t believe that will be the case for a second, but it’s worth noting that the numbers look bad for him right now.
• Kildare South: Just quick note on this as I was asked what will happen if the Ceann Comhairle retires instead of getting automatically returned, opening up a fourth seat. In all likelihood it will go to a second FG candidate, though Cathal Berry (IND) and a second SF candidate would be very competitive.
• Laois-Offaly: This is incredibly close for the final seat – not much would separate a second SF, a second FG, a second FF (last month I wrote off a second FF candidate; this was a calculation error on my behalf which has since been rectified) and Carol Nolan (IND).
• Tipperary: Tipp is getting increasingly hard to call – Matt McGrath (IND) is looking in much better shape, and FF, FG and Labour are all floundering. Candidate strategy is going to have such an enormous impact here and the projection here could change substantially if current trends endure.