So, just how many seats will they win?

By Michael Gallagher

With the election now just a few months away, possibly even closer, we should have a clear idea of how the parties will line up in the 31st Dáil. Nonetheless, there is evidently still a good deal of fluidity in voting intentions. Fianna Fáil support is rather like the banks’ liabilities: every time we feel that a bottom has finally been determined, there turn out to be further depths yet to which the graph can plummet. All that we have learned over the years tells us that FF cannot possibly poll as low as the 17 or 18 per cent at which recent surveys have recorded it, never mind the 13 per cent of last week’s poll, but this is the first election since 1927 that FF has gone into knowing it is doomed to defeat whatever it does, and much of what we have learned over the years does not seem to apply any more.

FG has shown no signs of a surge, and the decline in its vote in the Donegal SW by-election relative to 2007 is hardly a positive sign, but by-elections are different, and there is clearly a feeling that it has regained its traditional ascendancy over Labour, so at present it is in pole position to emerge as the strongest party.

Labour is still on course to achieve a record level of support, yet its fifth position in Donegal SW, the way in which its voting support in the 2009 local elections fell somewhat short of pre-election poll findings, together with uncertainty as to how its large number of new and relatively unknown candidates will fare when the voters come to mark their ballot papers, have all contributed to the continuing difficulty many people have in envisaging that it could emerge as the strongest party.

Sinn Féin, meanwhile, is also on course for a record haul of votes. Its first by-election victory since 1925 suggests that, unlike in 2007, its creditable opinion poll figures – over 10 per cent in several polls, up to 16 per cent in last week’s Red C poll – might actually translate into votes on polling day.

Finally, it’s worth observing that a striking absentee from the political scene is (so far, anyway) any new party. Confidence in the government is low, confidence in the opposition not a great deal higher (the Irish Times poll of late September found that only 39% thought a new government would improve the economy), but unlike the last crisis of the mid-1980s no new force has emerged to fill what might seem to be a sizeable gap in the market. In the other countries worst affected by the current economic difficulties – Greece, Portugal, Spain – too, voters are switching from some established parties to others rather than to new parties, a sign of the institutionalisation of many of Europe’s party systems.

If we knew how many votes the parties would win, how could we figure out the seat distributions? One way is to apply the party gains / losses in vote terms to each of the 43 constituencies. This has been done a number of times on these pages by Adrian Kavanagh and leads to some fascinating projections and discussions. There is a lot to be said, though, for the altogether simpler method of just applying the national vote shares to the total number of seats and generating estimated party seat shares that way. As the saying goes, if a party gets the votes it will get the seats, and if it doesn’t get the votes it won’t get the seats. Thus, if a party wins 20% of the votes, we know it will win about 20% of the 165 seats being contested (ie 33), even if we don’t know exactly where it will win these, which is an altogether more difficult conundrum to solve. Essentially, this assumes that parties will be over-represented in some constituencies and under-represented in others – a reasonable assumption, though not always true, as we saw in 2002 when FG lost out in more cases than it was over-represented.

Four other factors will affect the vote–seat conversion:

(i) party size. The largest parties invariably pick up a seat bonus just for being large, as their candidates stay in the count longest and are still around to receive transfers from smaller party candidates as these are eliminated. The small district magnitude employed in Ireland (only 4 seats per constituency on average) compounds this. This has always worked to FF’s advantage in the past, but at some estimates of its vote next time its benefit from this factor in 2011 will be small or even non-existent. And if SF support on the day reaches the level found by some polls, of 16% or so, it will benefit from this factor. All four main parties stand to receive some benefit, while, as always, Independents collectively will be the biggest losers from this.

(ii) candidates. It’s possible that some voters who tell pollsters that they intend to vote for, say, Labour, and indeed would do so if Éamon Gilmore was on the ballot paper in their constituency, won’t do so on the day when they find they know little about the local Labour candidates. The same might apply to SF; does it have enough strong candidates to capitalise on the rise in support for the party? Conversely, FF might hope that even if voters are now alienated by the party brand, some voters will still recall with gratitude the hard work of individual FF TDs over the years. The 2002 election study, after all, showed that many voters, according to their own accounts anyway, give more weight to candidate than to party when making their voting decision (The Irish Voter, chapter 8).
But, as against this scenario, experience shows us that when the tide runs strongly for or against a party, it affects it pretty much everywhere. The new votes for the PDs in 1987 and for Labour in 1992 swept into the Dáil candidates who were not well known even in their own constituencies, while FG’s big loss of support in 1987 saw many long-serving and hard-working TDs ousted. Likewise, while many voters might say that candidate rather than party is what determines their vote, we have to take this with a pinch of salt given that the candidate line-up at the next election won’t be hugely different from that in 2007 and yet we are expecting a very different pattern of votes. Indeed, not only will long-serving FF TDs not be protected by their record over the years, but longevity might even be a liability as voters express their rejection of people who they feel have been around too long and have contributed, if only through passivity, to the current state of affairs.

(iii) transfers. Parties that are everyone’s second choice can expect a higher return of seats per vote than parties that other voters prefer to keep at arm’s length. SF has always lost out from this in the past. The pattern of transfers in the Donegal SW by-election, where Pearse Doherty received nearly twice as many transfers from Labour as the FG candidate did, shows that the right candidate can overcome what is still, for many voters, a certain resistance to SF as a party, and transfers from smaller groups on the left may go in reasonable numbers to SF. The number of Labour transfers passing to SF is likely to be higher than in 2007, and though SF still will not pick up transfers from FF or FG in significant numbers, the number of FF transfers in particular is likely to be less than those from independents and smaller left-wing groups, so this won’t matter so much. For once, SF won’t be the main loser from transfers.
This time, FF will suffer most from the impact of transfers; it is now back to its pre-1997 position of having no allies and, moreover, being sufficiently unpopular to motivate supporters of all other parties to rank every other party above it. FG will not benefit much from transfers; Labour voters will prefer it to FF, but for groups further to the left FG is little more attractive than FF. Labour will be the main beneficiary of transfers. Groups to its left will prefer it to FF or FG; FG will prefer it to all other parties; even FF voters may rank it second, though many FF votes will become non-transferable once the last FF candidate is elected or eliminated. Overall, though, transfers between any two parties don’t look likely to be especially strong this time.

(iv) vote management. In the past this has been an issue only for FF and FG; other parties simply have not had enough votes to need managing. This time FF is likely to witness intense battles in many constituencies as two strong candidates, sometimes two TDs, fight for the (at most) 1 seat the party can win, with the risk that if things become too bitter, the internal transfer rate will become so low that neither candidate wins. FG has reasonable experience of the occasional need to manage votes, and that will stand it in good stead this time. The main question mark is over Labour, which according to the poll figures has a good chance of winning 2 seats in many constituencies, something that will in some cases depend on incumbents, who are accustomed to running alone and projecting themselves as much as the party, agreeing to share the votes equitably with a running mate. Time will tell how well the party adjusts to this new environment.

Putting these factors together, it is likely that for a given vote total Labour will do best in terms of seats, FG next best, SF third and FF worst. So, when the next opinion poll appears, to estimate the seat return for each party, give FF 1 seat (the outgoing Ceann Comhairle, assuming he stands again), and distribute the other 165 seats according to these principles:

(a) start by calculating the proportionate share of the seats that that share of the votes would lead to – eg if a party has 22% of the votes, 22% of the seats is 165 * 0.22, which is 36.

(b) for FF take a 2-seat range around the result – eg if it is shown with 19% of the votes, 19% of 165 is 31, add 1 for the Ceann Comhairle, leading to a prediction of 30–34; if it’s recorded with 26% of the votes the prediction would be 42–46.

(c) for FG add around 4 seats and take a range of around 2 seats above the result – eg if it is shown with 30% of the votes, 30% of 165 is 50, leading to a prediction of 52–56; if it’s on 36%, the range would be 61–65.

(d) for Labour add some seats (approx 2 if its vote is below 20%, 4 if its vote is in the 20s, 6 if it’s over 30%) and take a range of around 4 seats above the result – eg if a poll finds it has 21% of the votes, 21% of 165 is 35, giving a range of 39–43; if it’s on 28% the range would be 51–55.

(e) for SF add 1 and take a 2-seat range around the result – eg if it is shown with 10% of the votes, 10% of 165 is 16, leading to a prediction of 15–19; if it’s recorded at 16%, the prediction is 25–29.

Thus, if Labour and FF were each to win 25% of the first preference votes, we would expect FF to win not much more than 25% of the seats (around 40–44) while Labour could expect a larger bonus (around 45–49 seats in all). The best estimate, if we had to give a specific number, will be the mid-point of the range. There’s little point in trying to estimate the seats of smaller parties (Greens, United Left Alliance) or Independents, as at such small numbers local factors become more important.

For example, applying these principles to the results of the Red C poll published in The Sun last week (FF 13%, FG 32%, Labour 24%, SF 16%) gives a distribution of the order of FF 20–24, FG 56–60, Labour 44–48, SF 24–28.

Trying to convert votes into seat estimates is always something of an art rather than a science under any electoral system based on small constituencies, especially PR-STV, but national-level conversion is not only much more straightforward but is also likely to prove at least as accurate as making predictions for each of the 43 constituencies and totalling these.


14 thoughts on “So, just how many seats will they win?

  1. These figures (FF 20–24, FG 56–60, Labour 44–48, SF 24–28) seem more plausible that Adrian’s estimates (FF 12, FG 67, Labour 48, GP 0, SF 24, Inds/Oth. 15) given what we know about individual constituencies. Adrian is always clear that his model was not an exact guide, but all the potential ‘errors’ that were pointed out seemed to underestimate FF support, and so weren’t balancing out in the main.

    But are you being a bit conservative? Seat bonuses (and consequent penalty) can be much greater than you allow for – look at what the 5 point increase in support in FG between 2002 and 2007 translated into in seats. Of course it’s hard to be systematic on seat bonuses. Small parties with geographically concentrated support can get a seat bonus.

  2. Really interesting post – though another real problem is the variability of polls themselves, both over time and accross agencies.

    Probably this means that we realistically have to further widen our margins. Even assuming a (very generous) 3% margin of error for the Red C poll, this would widen our projection margins out considerably, taking across poll volatility into account we would probably do well to even further widen our margins.

    Also, polls in the run up to elections tend to under-estimate the vote share that the incumbent will receive elsewhere. I note that paddypower still has FF priced up as being most likely to win 30-40 seats, and this bracket seems pretty realistic to me.

  3. I think the estimated ranges on no. of seats by party is very useful. Taking the centre point of each range suggests that there is bloc comprising the populist, catch-all parties, FF & FG, falling just short of a majority. And there is a left-of-centre bloc – comprising some populist, nationalistic elements – also faling short of a majority. The usual ‘conventional wisdom’that FG + Labour will provide the alternative government may appear to have the weight of numbers, but both parties seem to be drifting further apart on the policy front. And FF is very cleverly highlighting this increasing divergence.

    My sense is that the FF vote will not fall as low as the range indicated. The likely muddle-headed nature of an FG/Lab combo, FF’s ability to highlight this, the Kenny factor, unease about the possibility of a Lab+SF+other left government and, yes, believe it or not – this is Ireland after all, a bit of a sympathy vote, certainly for the MoF’s personal circumstances but also for a government battered by the markets and the EU/IMF.

  4. 90 years of Fianna Fáil won’t be wiped out overnight – there are plenty who will vote FF no matter what. Even with this mess, I reckon FF will easily get 20%, how that works out seat wise and transfer wise is anyone’s guess.

    But there are also dangers in getting what you wish for as a government with a majority of 20 or more is not good either, especially in the country like Ireland, unless there is reform of the Oireachtas of such a scale, it would make the current occupants’ head explode, they’d find it so fantastical to comtemplate.

    The country will need a good opposition too as the new government will get things wrong but will FF or SF be up to that job? I really can’t see whoever survives in FF maturing enough to not revert to type and oppose everything for the sake of it, as they did between 1982-87 but then again this election is a game changer, more so than 1918 as we didn’t know in advance how much of a change that would mean – this time we know change is a coming. Time will tell …

  5. The most logical outcome of the seat distributions you project would be a FG/FF Government with c. 80 seats and Labour Sinn Fein opposition with 72. It would be a bit much to expect Labour, with 46 seats to act as second fiddle to FG with 58,and anyway a 114/48 FG-Lab/FF-SF Government/Opposition split doesn’t seem either very stable or desirable…

    However, I can’t see FF agreeing to act as second fiddle to FG either as this would only consolidate its position as a minor party. So we are likely to see a prolonged period of instability with no Government being agreed.

    The final outcome might be a FG/Lab Government with a rotating Taoiseach but you couldn’t rule out Lab/SF with support from virtually all the independents on the grounds that FF is too toxic for anyone to coalesce with (remember what happened Labour the last time) and the independents would have a lot of leverage with a minority Lab/SF Government.

    But what is extraordinary about all this is the degree of uncertainty of the Governmental outcome given the variability of polls, the variability of local constituency circumstances,and the relationships(or lack of) between the parties.Those who want certainty – e.g. the markets – will have to wait.

  6. I genuinely believe the most likely out come on these sort of figures would have to be a FG minority Govt. supported externally by FF. There is too many differences to form a stable Govt. with Labour and with them nearing 50 seats its likely they’ll push a larger amount of their agenda then was seen during the ’80s. Micheal Martin even admitted himself this week he was willing to support a govt. they would carry out a four year plan and went on about how FG proposals were much more credible. Also the idea FF would serve in Govt. as a junior partner is never going to happen and I don’t think FG want them there ether. We may be looking at a reverse Tallaght strategy here.

  7. As long as FG/L have the numbers they’ll form a government and they’ll have a year to do al lthe horrible stuff and blame on the FF/G/PD/Ind governments of the last 14 years.

    If the Tories and Lib Dems can form one there’s nothing to stand in the way of FG/L and the differences between them are nothing like as much as implied and you can be sure FG/L officals are working on the details at the moment – it’s crazy to think they’ll only start working out the details after the election and you might even find once the election is actually set, the two parties suddenly find they have more in common than they thought and both parties’ supporters will compromise on a lot – the alternative is more years in oppositions – no way.

    • Des fitz,i think you have it spot on, of course,in all coalitions, there will be disagreement, but one thing all non FF have in common, is get them out and keep them out.

      whatever the numbers are, labour will almost certainly be looking for a rotating Taoiseach, this in my humble opinion, would not be a bad idea.

      • The rotating Taoiseach is a non runner – if FG get the most seats they hold the Taoiseach and if Labour do then they get to be Taoiseach.

        LAbour being bigger than FG would be as much of a shock as whatever the FF vote will be. The heat of an actual general election will get rid of most of the fluff on the Labour vote – I expect the usual FG/L government in the usual proportion – the real worry is the problems that come with a huge majority but there is so much that needs to be done a productive job can be found for every single government TD. There won’t be any need for bored backbench TDs causing trouble – except for John Deasy of course but like father like son I guess.

  8. Yes, as Eoin OM says, the margin of error can’t be bounded definitively and as the example of FG in 2002 and 2007 shows, a party might gain (or lose) systematically rather than seeing its over- and under-representations even themselves out. But totals further from the mid-point of the range are less likely than numbers closer to that point.

    And, for clarification, perhaps I should reiterate that this is an attempt to estimate how vote percentages will translate into seats, not an attempt to predict what the various parties’ vote totals will be.

  9. Pingback: Fianna Fail: The meltdown is fact! « Slugger O'Toole

  10. The Irish Times poll published today would suggest something of the order of FF 29 seats, FG 54, Labour 47, SF 26 on the basis of the rationale outlined above.

    That’s probably a bit lower than most people expect for FG – but remember that FG did unusually well in converting votes into seats in 2007 and this will not necessarily happen again.

    It’s certainly much higher than most people expect for SF – Adrian Kavanagh’s projection from the same poll, based on constituency-level predictions, suggests 18 seats for SF. But if SF actually does receive 15% of first preference votes, despite having many relatively not well known candidates, then, given that for a change it can expect to do reasonably well in receiving vote transfers from minor parties, Independents, and perhaps some Labour voters, most of the reasons why in the past its share of first preferences have translated into a markedly smaller share of seats won’t apply this time.

  11. Pingback: Local polls may be best guide to seat destinations «

  12. Pingback: Poll of polls: weighted average of party support «

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s