March 2022 Projections Update

Seat Map

Three polls this month, as usual B&A and RedC – and an Ireland Thinks with no breakdown, so no application to the model – but with a significant gap on Fine Gael. RedC had them at 19%, which is the lowest they have had them since at least 2006 (the furthest back I could find). On the other hand, B&A had them at 24%, the highest since July 2021. This is quite a contrast, and more typical of the polling gap between the two we see for Fianna Fáil. There is, however, a clearly identifiable reason for the gap on FG, which I’ll discuss below.

FG, it must be noted, are clearly spooked by that 19% figure, and as usual anonymous backbench TDs are complaining about the party leadership. The FG leadership, for their part, have responded by making an increasingly unrealistic series of tax and spend proposals (i.e. no taxes and loads of spending). Never let them tell you the magic money tree isn’t real!

Overall however, we’re not seeing any dramatic movement, just the continual pressure on FG. Sinn Féin seem to be bumping up against another cap in support at around 33%, but we’ve seen them flatten out before moving up in the past – they hovered around 28% for a long stretch in late 2020 and early 2021, and again around 30% in mid-2021. The result of this is on seat projections is some swapping around between FF and FG, as well as some other candidates coming back into the mix as FG struggle. However, the trend is culminating a bit, as I’ll talk about in the first section below.

There are quite a few additional areas I’m going to cover below, as there’s been some interesting stuff going on this month, but if you would prefer to skip straight to the seat changes , click here.

The government’s seat total is… bad

The projections this month have the combined government (Fine Gael, Fianna Fáil and Green Party) total at 75 seats. That’s the lowest I’ve ever had it at. Although not the first time dipping below a majority, 5 seats below puts them in a really awkward (and, to me, funny) situation – and that increases to 6 seats, if they wish to hang onto having the Ceann Comhairle. There have been CCs before from non-government parties (most recently in 1997-2002), so ditching it could be viable, if not preferable.

But then how do you make up this gap, be it five or six seats? The most likely possible partner is whatever is left of the Labour party, and by our modelling, that’s not going to be enough (and if Labour did win more seats, a number of those would be at the expense of FF or FG, and possibly a Green, so it wouldn’t change the overall calculus). At that point you are looking to cobble together a bunch of random Independents to hold things together.

That’s not a recipe for stable government, nor is it a good sign for advancing whatever your agenda is if, say, Michael Healy-Rae, can throw his toys out the pram and threaten to bring the government down every time he doesn’t get his way. While well-meaning centre-left parties have some incentives to put up with being disrespected by larger partners, the dynamic with individual operators with secure seats is very, very different.

Basically, this is not a good position for the government to be in, polling-wise, and unless things substantially improve across the board, the chances of it being viable following an election continue to diminish.

Strange polling in Dublin for FG (and the SDs)

I mentioned further up the gap between RedC and B&A on Fine Gael – around 5%, and that Fine Gael seem shaken by the 19% poll and not reassured by the 24% one. Now, this may be in part because it suits the agenda of those who are unhappy with the current party leadership, but I actually think they’re right to put more weight on the worse poll.

There’s a simple reason for this – the B&A poll having them on this number is due to an colossal outlier in Dublin, where it has Fine Gael at 36%. Fine Gael haven’t been polled this high in Dublin since February 2021, and that was a RedC poll. B&A haven’t had them that high since before GE 2020 (they had them at 35% in October 2020). That 36% figure is over 10% higher than their recent polling in Dublin, and has come out of nowhere.

B&A do have a habit of pulling out really strange numbers that seemingly nowhere – we saw a period where they had Labour very high in Rest of Leinster (and the same for Aontú right now), and of course there’s the huge gap they have in polling Fianna Fáil against RedC. But these do tend to endure beyond a single poll, and it doesn’t neccessarily mean they are wrong. It’ll be interesting to see if they have a similar figure next month, or if this is a true outlier.

Either way, it makes that 24% poll, for now, out of place and generally questionable against what we have seen in prior polls, and the 19% is more in line with recent trends. This is why polling averages are great – they smooth out things like this.

Speaking of, RedC have suddenly got the Soc Dems on 11% in Dublin. The SDs have been doing well in RedC in the capital of late, tracking around 7% in recent polls, though B&A have them much lower. But this is a very sudden increase and again tracks far ahead of recent polling. The same caveats as above apply, and it should be considered an outlier until we get more data showing it isn’t.

TCD Seanad By-Election

This section was originally just going to say “I don’t care”, which was part of the reason I made no effort to do a projection on this (the other reason was the lack of data). But then the count started, and, well, I have a very unhealthy relationship with elections, as you can tell. I’ve done some surface-level transfer analysis here, and will probably do a write-up next week.

Transfers

Short note on this: as the polls move more and more away from the 2020 base, I’m seeing stranger things in the transfer system. This needs a re-work, and I’ve flagged this before, but it is a substantial piece, and has already been re-worked before. If people have thoughts on how this could work better (or on how modelling Independents could work better), please let me know!

Limits of the model and local expectations

A final note that I find interesting. The system used for seat projections will always be imperfect absent constituency level data, making assumptions based on regional polling. Some of these look wrong and will be wrong. Others look wrong and might not be – we won’t know until an election happens. I think there’s an interesting illustration of this in Munster with Sinn Féin. The polling average has them increasing their vote in the province by 14%, which is huge (a +73.3% swing).

The model we use here has to assume – as it has no better data – that this is relatively evenly distributed across constituencies. Now, in reality, it won’t be, and that will result in some inevitable misses. I’ve been keeping track of where people have told me Sinn Féin won’t gain a seat:

• Clare (model agrees now, but hasn’t in the past)
• Cork East
• Cork North-West
• Cork South-Central
• Cork South-West
• Kerry
• Limerick County
• Tipperary (model agrees)

The funny thing is that, taken in isolation, all of these have reasonable arguments for being correct; there are factors that could prevent Sinn Féin from gaining a seat in any of them.

But taken together, it leaves Cork North-Central, Limerick City and Waterford (where the model has SF on three seats, which is I agree questionable) – a total pick-up of three seats, despite a vote increase of 14% to 33% overall. That’s borderline implausible (and would indicate a disastrous failure of either polling or of electoral strategy), unless the new SF vote is incredibly concentrated into a handful of constituencies.

I suppose the point here is that there are limits to any system of projection or prediction, and an open mind should be kept. People in general do know what they are talking about, and data is useful, but both have limits. Elections, at constituency level at least, can and do always throw up surprises.

Seat Changes

Six constituencies saw changes in March; including two changes in Dublin from last month reversing.

Clicking on the linked name for each constituency should jump to the relevant section of this page. Changes on this page indicate changes from February 2022’s projections; changes on constituency pages indicate changes from current composition.

Note: The projections reflect, and always have reflected, most likely outcomes. So if a final seat is more likely for candidate X over Y, the model will show X winning the seat.

This does not mean the scenario where Y wins doesn’t exist, or even is necessarily unlikely (there’s a lot of marginal calls!). It also does not mean that every single “most likely” scenario will come true; statistically that in and of itself is probably not going to happen. This is true from from a simple probability point of view, even if we ignore deficiencies in underlying data. A projected result merely means that the model thinks X winning is the most likely outcome.

• Cork North Central (PBP +1, FG -1)
• Donegal (IND +1, FG -1)
• Dublin Bay South (FG +1, FF -1)
• Dublin Central (SD +1, GP -1)
• Dublin Mid-West (SF +1, FF -1)
• Sligo-Leitrim (IND +1, FG -1)

This is an odd one, and part of the reason I think how the model currently handles transfers needs to be revisited. This would be a huge result for the left – two Sinn Féin, and Mick Barry holding the Solidarity seat in the constituency. There’s an awful lot of contingencies here, mostly around how an SF vote split would pan out in reality, but also on transfers and of course whether or not ex-Fianna Fáil Independent Cllr. Ken O’Flynn runs again (the model currently assumes he will).

The key factor driving this is an increasingly problematic situation for Fine Gael in Munster. They are down over 3% (a 14.7% swing) on their 2020 result, and are way behind FF – moreso than they were in 2020. There’s also the volatility of PBP/Solidarity polling in Munster, which is very low and thus highly susceptible to wild swings based on relatively small polling movements.

What I’m ultimately getting at here is that this may well be illusory; there’s a lot of factors going into how the model is reaching this conclusion and a shift in any of them will probably flip this seat back. However, even if that is the case, Barry is right now in a more competitive position than he’s been since I started this project.

(And yes I know Mick Barry is technically Solidarity, not PBP. I’ll change the notation if the relationship between the two changes and they do not run as a joint party.)

Before this month, Donegal hasn’t changed since I started doing these projections, but I’ve mentioned prior that Thomas Pringle (IND) has been creeping closer and closer to those ahead of him, and now it reckons he holds the edge over Fine Gael’s Joe McHugh, though there’s quite a few factors here that make this fuzzy.

Firstly, there’s the ongoing fallout of the mica scandal and the failure of the government to address it. This will doubtlessly impact Fine Gael, but Fianna Fáil’s Charlie McConalogue has also come in for a lot of criticism over it, and while the model has him holding his seat with a high degree of probability, this is exactly the kind of constituency nuance that a model based on provincial level data could very easily miss.

Secondly, the model assumes Sinn Féin running three and getting three. They probably could have pulled this off in 2020 at Pringle’s expense, but running an effective vote-split strategy with three candidates is not easy, and how that happens and how their surpluses break down is tricky to say with great certainty.

Thirdly, there’s a massive discrepancy between Red C and B&A on how Fianna Fáil are doing in Connacht-Ulster; while the model smooths this out, if either one is correct, it will significantly change the dynamic for them and for the candidates around them.

There’s also, of course, the possibility that the mica scandal itself breeds another candidate to run on that issue. At that point, all bets are off.

Last month I described Dublin Bay South as “extremely messy and close” and this month has borne that out, with the final seat flipping back from Fianna Fáil to Fine Gael, but it’s all very fine margins stuff. Of course, the model can’t factor in stuff like incumbency effects, which, in theory, should favour O’Callaghan (FF), but honestly, who knows?

It’s a shame that after the projections for Dublin Bay North have got a bit more stable, the chaos has just taken a short jaunt south across the bay, and this could well move back and forth while FF and FG remain close to their current levels of support. Suffice to say this isn’t something that could be called with any great confidence.

Not a lot to note here, as the change from last month flips back. The SDs got a crazy high result in Dublin in the B&A poll this month (11%!), and while the RPA has successfully smoothed that out for the most part, it has had an impact in this constituency, where a razor-thin margin continues to separate the Social Democrats, the Greens and Fianna Fáil in the race for the final seat. With FF and the Greens generally trending upwards in Dublin, I suspect this will remain competitive for a while yet.

Who gets eliminated first will be key. If FF or the SDs are knocked out first, that would favour the Greens; if the Greens go first, that would favour the SDs – essentially making it quite difficult for FF to get ahead. The presence of a second Sinn Féin candidate is, however, a bit of a wildcard when it comes to transfers, so there’s still a lot of unknowns here.

This is the first time, I think, that the model it has shown Sinn Féin getting three seats in Dublin Mid-West. Sinn Féin are a bit stagnant in Dublin again, but Fianna Fáil drop a bit after the most recent round of polling, and that reflects here. But this is still a bit strange and might just be one of those temporary things where we just happen to be at a particular point in the polling movement of the parties, rather than a trend culminating.

With that said, it’s not completely implausible. Sinn Fein got 42.8% – 2.14 quotas – last time out, and the model does expect them to improve that here. Given how the rest of the constituency shook out, there are realities, albeit marginal ones, where in 2020, they have enough votes for three seats. Naturally, all this depends on quite a few contingencies around vote split and transfers, which is itself cause for caution – because SF dominated the vote here, both Ó Broin and Ward were elected before we could see if anyone else would transfer to them.

Also, I should note, I don’t think Sinn Féin will actually end up running three candidates here. It’s risky, despite what the maths might say. If they do only run two, given how their surplus will go, I’d expect Gino Kenny (PBP) to be the main beneficiary.

Polling for Independents in Connacht-Ulster has been a bit better lately, whereas Fine Gael’s numbers have been pretty awful; that reflects here as the model reckons Marian Harkin should be favoured to keep her seat. However, the usual caveats should be applied when considering Sligo-Leitrim: we don’t know what Marc MacSharry will do, or how what he does will impact the end results, and in 2020 there were literally nineteen candidates, which created substantial vote fragmentation. So there’s a lot that can happen here that we can’t account for with a ton of confidence.

Other notes

• Dublin Bay North: Feels redundant to point this out again, but the probability gap between Labour and the Soc Dems is microscopic once again and the most likely outcome could change for the hundredth time shortly.
• Kerry: Closest Danny Healy-Rae has been shown to holding his seat for a while, but this is more a function of Fine Gael’s weak polling than anything else.
• Kildare South: Small note here, with the Labour vote in Rest of Leinster receding in the polls, if the CC retires, Mark Wall is no longer favoured to take the fourth seat; instead it looks more likely that Sinn Féin could get a second candidate in.

Thanks so much for reading! This website is done entirely in my spare time and run without ads, so if you want to donate, please do so via Patreon or Ko-Fi – any support is greatly appreciated and 100% of any donations received are invested into the costs of running and improving the site.