How parties' poll positions might translate into seats
There is a way to calculate the chances of a vote share securing the election of a TD, writes MICHAEL MARSH
THE IPSOS MRBI /Irish Timespoll yesterday provided further confirmation of all the trends which are becoming apparent in this campaign: a clear increase in Fine Gael support, a decline in Labour’s fortunes and no significant change elsewhere. Fianna Fáil remains in the doldrums, Sinn Féin has little hope of a major breakthrough this time and Independents are remarkably popular.
The relatively small – but highly significant – changes during the campaign should not blind us to the reality that this election will establish some new records and equal other existing ones if the results match what is found in this poll.
Fianna Fáil’s 16 per cent would be its lowest figure to date, Labour at 19 per cent could equal its Spring tide of 1992, 11 per cent for Sinn Féin would be that party’s best performance in its current incarnation, while 14 per cent for Independents would arguably constitute the biggest non-party vote yet. At 37 per cent, Fine Gael is only matching its second-best figure, achieved in 1981, but, with the wind apparently behind it, the 39 per cent of the vote won by Garret FitzGerald in 1982 is well within reach.
What does this mean in terms of seats? There are many ways of predicting how these vote shares will transmute into TDs.
Prof Michael Gallagher of Trinity College Dublin has argued (on electoralreform.ie) that the most reliable way we can translate these numbers directly into seats is by assuming we really do have a system of proportional representation, then making a few adjustments on the basis of recent historical record – the largest party tends to get a large bonus, Independent votes typically translate poorly into seats – and possible transfer receptiveness.
Of course this bonus varies: Fianna Fáil came very close to an overall majority with almost 42 per cent in 2002, but was further away in 2007, despite a slightly higher vote. In 1997 the party won a single seat fewer than in 2007, but did so with 2 per cent less of the vote. This is part of what Brian Lenihan snr, musing on his defeat in the presidential election, called the “vagaries of PR”.
An alternative is to calculate what the national figures might mean at constituency level by making assumptions about constituency-level swing, then examining the entrails of each contest and totting up likely winners. This is what Noel Whelan has done in these pages, while Dr Adrian Kavanagh of NUI Maynooth calculates for each published poll for electoralreform.ie.
It requires a lot of constituency-level knowledge and disinterested judgment. Such calculations are also risky as swing is not uniform. In this campaign, polls suggest that the swing to Fine Gael is very pronounced in Munster and Connacht-Ulster but has not happened at all in Dublin. Even where a party’s support at national level remains the same at successive elections, as did Fianna Fáil’s in 2002 and 2007, it will rise quite significantly in some areas and fall significantly in others.
Indeed, although predictions for 2007 were relatively accurate for Labour and Fine Gael, the consensus was quite wide of the mark for Fianna Fáil and Sinn Féin.
I have explored a third way, working with Kevin Cunningham, a research student in political science in TCD. We have calculated the chances of a particular vote share being translated into a seat, using the last few elections as the basis for estimating the odds. In other words, if a party wins 60 per cent of a quota, what are the odds, based on past electoral records, that it will win a seat? Do those change if it has two candidates? And are the odds different for each party?
We can apply these odds to what the polls suggest will be constituency vote shares next Saturday. This entails estimating constituency support for each party using a poll, or series of polls, but taking the regional breakdowns rather than the national figure, as that promises a little more accuracy – although it certainly does not avoid the problem that swing is not constant.
As might be expected, the chances of a party translating votes into seats varies by party – Greens have historically done better than Sinn Féin. This is due to the relative ability of each party to attract second and later preference votes from supporters of other parties. The ability to convert support into seats is also dependent on the number of candidates a party nominates.
There is a clear cost of over-nomination in this regard. To see how many seats a party can be expected to win, we can then sum the probabilities of it winning a seat (or two, three or even four seats) across all constituencies.
One advantage of this approach is that it suggests that a party with a 33 per cent chance of winning a seat in each of three constituencies will – on average – get one seat. We are not assuming the most likely outcome in each constituency will always come to pass. This also means that variable swing is less likely to upset our overall calculations.
On this basis, and taking yesterday’s poll as the basis for estimating votes next Friday, our analysis is that Fine Gael will win 72 seats, Labour 35, Fianna Fáil 26, Sinn Féin 11, Greens two and others 19.
As with all such estimates, there is a defined band here of a few seats either way, so the Greens could very easily get no seats. What may prove significant in this election is the dramatic change in the attractiveness of each party to voters casting a second preference.
The Greens have done well out of PR-STV in the past and were able to win seats despite getting less than half a quota on first preferences. Some evidence suggests that this time the party will be less popular.
Yesterday’s Irish Timespoll asked about alternative preferences. If we take this as indicating probable transfer patterns, responses did seem to indicate that the Greens might not be completely toxic. The same responses did point to serious problems for Fianna Fáil, as only 7 per cent said they would also consider that party, compared with 12 per cent for Fine Gael and a whopping 23 per cent for Labour.
Fianna Fáil is in severe difficulties in many constituencies if the long-expected poll bounce continues to elude Micheál Martin, a problem exacerbated by his failure to correct over- nomination. There are 32 constituencies where it seems unlikely to achieve a full quota on first preferences alone and only one, Carlow-Kilkenny, where it might be expected to get two quotas.
Labour is unlikely – on these poll results – to reach a quota on first preferences in 25 constituencies, but in contrast to Fianna Fáíl, it is likely to pick up support from all other parties as they are eliminated, being the second favourite party of a majority who acknowledge such a preference. This is particularly true of Independent voters, 30 per cent of whom opt for Labour as their second preferred party, as opposed to 14 per cent for Fine Gael and a miserable 4 per cent for Fianna Fáil.
Twenty-six per cent have no other (party) inclination. Even if Labour’s vote continues to fall and that party ends in third place in terms of votes, it remains likely to be the second largest party in the next Dáil.
There is no obvious indication in the Irish Timespoll that Fine Gael will attract transfers much better than it has in previous years. However, to the extent that this contest looks increasingly like one between a big party and three much smaller ones, the big party could well get an unusually large bonus as it stays in most races until the end.
Fine Gael could fall short of a quota in only one constituency (Kildare South). It seems likely to get more than two quotas in nine seats and has a good chance of winning two seats in a majority of constituencies (31) and more than that in a few of them.
There is also the possibility of a remarkable 3.8 quotas in Mayo, leaving aside what might be the impact of Enda Kenny’s elevation to taoiseach on Fine Gael support in this particular constituency.
Fine Gael’s chances of an overall majority still look remote on the basis of this poll alone, but the growth seen in this one over the previous Ipsos MRBI/Irish Times poll is on a par with that shown by successive surveys conducted by other companies. Taken together, these suggest the party is picking up support steadily, of the order of 1 per cent every four or five days.
Polls published last weekend were conducted at least a week before polling day so, if the trend continues, this could mean Fine Gael would be almost 2 per cent higher next week, putting it somewhere – according to different polls – between 39 and 41 per cent. Anything in this range will ensure that public interest in the count will stay high right through Saturday and well into Sunday.
Michael Marsh is professor of comparative political behaviour and pro-vice-provost of TCD