Adrian Kavanagh, 25th June 2012

The Sunday Business Post-Red C (24th June) poll notes a recovery in support levels for Fine Gael, with Sinn Fein support levels falling relative to the party’s high level in the 26th May poll.  The poll puts national support levels for the main political parties and groupings as follows: Fine Gael 32% (+2%), Labour 15% (NC), Fianna Fail 16% (NC), Sinn Fein 16% (down 3%), Green Party, Independents and Others 19% (up 1%). In line with previous posts which have applied a constituency level analysis – using the constiuency units used in the 2011 General Election – based on assigning seats on the basis of constituency support estimates (simply using a d’Hondt method to determine which party wins the seats), while also taking account of the factors of vote transfers and vote splitting/management (based on vote transfer/management patterns oberved in the February 2011 election) – party seat levels, should such national support trends be replicated in an actual general election, would be estimated as follows: Fine Gael 66, Labour 26, Fianna Fail 28, Sinn Fein 23, Green Party, Independents and Others 23.  However, if the new constituency boundaries, drawn up in the recently published 2012 Constituency Commission report, were to be used instead as the basis for the constituency level analysis, the party seat numbers would now be estimated as follows: Fine Gael 67, Labour 21, Fianna Fail 27, Sinn Fein 21, Green Party, Independents and Others 22. This analysis suggests that Labour would seem to be the party most likely to be adversely effected by the redrawing of the constituency boundaries and suggests that the party, in this context, would suffer the main impact of reduction in Dail seat numbers from 166 to 158, while Fine Gael would be seen to gaining somewhat from Labour’s misfortunes and actually gaining a further two seats arising from the boundary redrawal.

The constituency support estimates based on the poll figures, when using the old constituency units (as used for the general election of February 2011), are as follows:

Based on these constituency estimates and using a d’Hondt method to determine which party wins the seats in a constituency, the party seat levels are estimated as follows:

Amending the model to account for seats that may be won or lost on the basis of a large/small number of candidates contesting the election (e.g. Others being allocated a seat in Laois-Offaly mainly due to the large number of independent candidates who contested this constituency), vote transfers and vote management (e.g. discrepancies between votes won by party front runners and their running mates which would see potential seat wins fall out of a party’s hands), the seat allocations across the constituencies would look more like this:

But the boundaries used for the 2011 election will not be used at the next general election, unless a snap election is called before these new general election constituencies are legally enacted by a new Electoral Act, and an constituency analysis based on these as such is of limited use in terms of trying to estimate seat numbers given the context of new general election constituency boundaries. In light of this I have attempted to re-create this constituency level analysis while using party support figures for the last general election based on the areas covered by these new constituency units and not the old constituency areas. Where tally figures are readily available (as is the case for most of the constituencies in the west of Ireland, as well as a number of other rural constituencies, thanks to the tradition of publishing the election tallies in local newspapers in the weeks follow a general or local election), a fairly accurate estimate can be gleaned of what the General Election 2011 party support levels for new constituency units would have been. In other cases, where tally figures are not available, a best-guess estimate of party support levels has to be employed, or else figures for the current constituency unit are employed. For instance the base figure for Dublin Bay North is simply taken as the total for the old Dublin North Central and Dublin North East constituencies as the lack of tally data for these constituencies means that one cannot estimate the impact of areas such as Portmarnock being moved into neighbouring Dublin Fingal. Similarly, a lack of tally information means that the base figure for Dublin Central has to be taken as that for the old 4-seat constituency unit although the loss of the Drumcondra area to Dublin North West and the Ashtown area to Dublin West would suggest that the figures for Fine Gael and Fianna Fail may be significantly over-estimated. This is not as scientific as one would wish it to be, arising from the lack of tally data for a number of the constituencies especially in the Dublin and South Midlands regions, and some constituency estimates will not be as reliable as is the case for the others but this analysis does give a sense of what the new Irish electoral geography following on the 2012 Constituency Commission report, might look like. Hopefully, if and when I can access tally information for other constituencies, the underlying data set will become somewhat more reliable in further such analyses.

The constituency support estimates based on the poll figures, now using the new constituency units (as defined in the 2012 Constituency Commission report), are as follows:

Based on these constituency estimates and using a d’Hondt method to determine which party wins the seats in a constituency, the party seat levels are estimated as follows:

Amending the model to account for seats that may be won or lost on the basis of a large/small number of candidates contesting the election, vote transfers and vote management (e.g. discrepancies between votes won by party front runners and their running mates which would see potential seat wins fall out of a party’s hands), the seat allocations across the constituencies would look more like this:

In both analyses, the slight increase in Fine Gael support estimates (increasing by two percentage points) and the stabilisation of the percentage Labour support level around the mid teens bodes well for a return of the government parties to power should these support levels be replicated at an actual general election. Should the seat estimates based on The Sunday Business Post-Red C poll figures pan out after the next general election, the the government parties would continue to hold a majority (albeit a somewhat reduced one, largely due to a significant reduction in Labour seat numbers) in Dail Eireann with seat numbers for a Fine Gael-Labour coalition estimated at (i) 92 seats in the old 166-Dail seat model, giving a potential coalition involving these parties a relatively comfortable majority of 18 seats in the Dail or (ii) 88 seats for the new 158-Dail seat model, giving a potential coalition involving these parties a relatively comfortable majority of 18 seats in the Dail. It is interesting that the size of the potential government majority remains the same when both the old constituency boundaries and the new constituency boundaries are used as the basis for this analysis, although the relative strengths of the two government parties in terms of seat numbers are seen to vary somewhat. Seat numbers for an alternative Sinn Fein-Fianna Fail coalition alliance are estimated at 51 seats for the old 166-Dail seat model and at 48 seats for the new 158-Dail seat model. Given the fact that a Fine Gael-Sinn Fein coalition would be highly improbable, the only other likely two-party coalition option would involve Fine Gael and Fianna Fail, with these parties’ combined seat levels based on these figures estimated at 94 seats for both the 166-Dail seat and the 158-Dail seat models, giving a potential coalition involving these parties a majority of 22 seats in the Dail for the 166-Dail seat model and of 30 seats in the Dail for the 158-Dail seat model.

The changing seat numbers for the parties in the different analyses points to one reason why the Irish electoral system is not entirely one hundred percent proportional – thus underpinning the rationale behind this series of constituency level analyses of polling figures – as the manner in which constituency boundaries are drawn, or redrawn, can act to gives certain parties a significant advantage in terms of translating their vote tallies into seat wins. This knowledge, of course, formed the basis for the gerrymanders that marked the partisan boundary redrawal system which existed up to the 1977 General Election, with the party/parties in government being in a position to be able to redraw election boundaries in a manner that would allow them to pick up extra seats. In simplistic terms, in the 1960s and 1970s this amounted to the main government options seeking to create constituency units with odd-numbers of seats in the regions of the state where their support levels were highest (where a 50% share of the vote would be sufficient to allow them win 2 seats in a 3-seat constituency or 3 seats in a 5-seat constituency) and constituency units with even numbers of seats (i.e. 4-seat constituencies) where their support levels were weaker as a 40% share of the vote would be sufficient to allow them win 2 seats out of 4. Since the introduction of independent boundary commissions following the 1977 General Election, partisan influences no longer can skew the boundary drawing process in favour of a government party, or government parties, but as the example here shows a significant redrawal such as that envisaged in the 2012 Constituency Commission report will probably tend to disproportionately advantage, or disadvantage, certain parties or political groupings. Similarly, as the range of constituency level analyses prior to the 2011 General Election displayed, a party’s ability to take advantage of such disproportionality in the system, whether arising from constituency boundaries or a tendency for the Irish system to favour the larger parties, is dependant on that party maintaining its support at, or above, a certain level, as a fall in support for that party, even if relatively minimal, can lead to disproportionate level of potential seat losses if party support levels fall below a certain “tipping point”.