Adrian Kavanagh, Sunday 25th September 2011

The main message coming from the two Red C-Sunday Business Post opinion polls, published on 25th September, is that vote transfers would see Michael D Higgins win the presidency despite trailing David Norris in terms of first preference votes. In all, the polls offer mixed messages for the government parties, some cold comfort for Fianna Fáil in the wake of the previous week’s disastrous Millward Brown Lansdowne poll figures for that party, and very good news for Sinn Féin and the Others grouping.  Transfer figures provided with the presidential election poll suggest that, in order to go on to win the contest, Davis Norris would need to almost ten percent ahead of Michael D Higgins on the first count, while Martin McGuinness would need to be eight percent ahead.

The Red C-Sunday Business Post poll puts support for the five confirmed candidates and two other potential candidates in October’s presidential election contest as follows: Gay Mitchell 13%, Michael D. Higgins 18%, Sean Gallagher 11%, Mary Davis 13%, Martin McGuinness 16%, David Norris 21%, Dana Rosemary Scallon 6% with second preferences expected to go as follows based on the poll figures: Gay Mitchell 14%, Michael D. Higgins 20%, Sean Gallagher 14%, Mary Davis 16%, Martin McGuinness 8%, David Norris 6%, Dana Rosemary Scallon 7% Based on these poll figures, I would predict that Michael D. Higgins will win the election, beating David Norris on the final count by 524,593 votes to 410,264 votes.

First step in this model will be to estimate the turnout for the election. The turnout is likely to be lower than the turnout for the general election, but there will be likely to be more people voting than in the previous presidential election in 1997. Fortunately there was also a general election in 1997 and this can offer a yardstick to help guesstimate the turnout. The numbers turning out to vote increased by 24.1% between the general elections of 1997 (1,806,932 voting) and 2011 (2,243,176 voting). Applying the same level of increase to the number that turned out to vote in the presidential election of 1997 (1,279,688) gives us a turnout of 1,588,641 voters. Given that 1997 was a highly uncompetitive election in which Mary McAlese had established an unassailable poll lead some weeks before the election, the likelihood is that the high profile afforded the current contest added to the very competitive nature of the contest suggested by these, and other, poll figures, the likelihood is that the turnout may well be higher than what is estimated here. Indeed the possibility of a record turnout for a presidential election, surpassing the 65% level for the 1966 contest, cannot entirely be discounted.

Based on the poll figures (for support and transfer patterns) and this estimated turnout value, the counts would be likely to progress as follows:

	1st	2nd	3rd	4th	5th	6th
Norris	340,423	350,149	359,308	361,863	390,910
Higgins	291,791	312,217	367,171	418,270	476,362	581,908
McGuinness	259,370	272,987	302,296	322,735
Davis	210,738	236,999	268,140	329,459	394,006	433,097
Mitchell	210,738	224,355	255,496
Gallagher	178,317	183,180
Scallon	97,264

There is some glimmers of good news in this poll for all the seven candidates/potential candidates listed here. While the addition of Norris as a candidate would see him lose the front runner status he has enjoyed in recent weeks (and also lose roughly one third of the first preferences votes he could expect to win should Norris not contest), The ability of both Higgins and Davis to attract transfers over and above any of their main rivals would see them edge out their current main rivals, Norris and McGuinness, as the election count progresses, leaving both candidates to contest the final count (with Davis narrowly edging ahead of Norris on the penultimate count). Norris will be heartened by his strong poll showing, as will McGuinness, but weak transfer figures for both candidates suggests that they both need to increase their first preference shares significantly over the campaign if they are to win the race for the Aras. Mary Davis is emerging as the dark horse in this election based on  her first preference and especially her ability to attract vote transfers, while neither Sean Gallagher or Gay Mitchell are out of contention and are well placed to allow a strong showing by either to push them back into serious contention, especially given their relatively stronger transfers ratings relative to Norris and McGuinness. As for Dana, while this poll suggests she is well off the pace, this poll figure is relatively similar to the levels she won in polls during the 1997 campaign, and these polls tended to significantly under-estimate her support as the final election figures showed.

The main message coming from this poll perhaps relates more to the findings of the transfer patterns, especially given the closeness of the contest as suggested by the poll findings relating to first preference figures. The low figures for transfers associated with McGuinness and especially Norris suggests that these two candidates will need to be polling significantly higher in terms of of first preference tallies than more transfer-friendly candidates, such as Davis and Higgins, if they are to go on to win the contest. So how far ahead would Norris and McGuinness need to be on the first count, based on support patterns evident in these poll figures, in order to go to be the winners once all the counting has concluded? First looking at David Norris, by replicating the model while increasing his vote and keeping the relative poll ratings of the other candidates consistent we can see that for him to edge out Michael D. Higgins narrowly on the final count that he would need to be over thirteen percent ahead of his closest rival on the first count – these figures would have the candidates’ support levels as follows – Norris 29.3%, Higgins 16.5%, McGuinness 14.7%, Davis 11.9%, Mitchell 11.9%, Gallagher 10.1%, Scallon 5.5%:

	1st	2nd	3rd	4th	5th	6th
Norris	465,153	473,908	482,151	484,451	510,595	528,326
Higgins	262,634	281,018	330,481	376,473	428,761	528,058
McGuinness	233,452	245,708	272,088	290,486
Davis	189,680	213,317	241,346	296,537	354,634
Mitchell	189,680	201,936	229,965
Gallagher	160,498	164,875
Scallon	87,545

Carrying out the same analysis in the case of Martin McGuinness shows that he would need to be nearly eleven percent ahead of Michael D Higgins on the first count, based on these transfer patterns, in order for him to enjoy a (very!!!) narrow win on the final count – McGuinness 26.9%, Norris 18.7%, Higgins 16.0%, Davis 11.6%, Mitchell 11.6%, Gallagher 9.8%, Scallon 5.3%:

	1st	2nd	3rd	4th	5th	6th
Norris	297,220	305,712	313,709	315,939	330,322
Higgins	254,760	272,593	320,573	365,187	445,728	534,915
McGuinness	428,068	439,956	465,546	483,391	492,021	534,963
Davis	183,993	206,922	234,110	287,647
Mitchell	183,993	195,882	223,071
Gallagher	155,687	159,933
Scallon	84,920

The extent to which both candidates’ share of first preferences would need to increase would not be as significant if their main challenger was a candidate such as Gay Mitchell, lacking the same ability to attract significant transfers to them as Michael D Higgins would. If Norris ended up vying for the presidency with Mitchell (in this case replicating the above analysis but swapping the positions of Mitchell and Higgins) he would need to be just under eight percent ahead of what would be his main challenger on the first count in order to win – Norris 25.2%, Higgins 12.6%, McGuinness 15.5%, Davis 12.6%, Mitchell 17.5%, Gallagher 10.7%, Scallon 5.8%:.

	1st	2nd	3rd	4th	5th	6th
Norris	400,093	409,354	418,076	430,842	457,437	501,120
Higgins	200,664	220,113	272,440	343,930	397,121
McGuinness	246,971	259,937	287,845	295,504
Davis	200,664	225,670	255,322
Mitchell	277,842	290,808	320,460	374,078	397,718	500,970
Gallagher	169,793	174,423
Scallon	92,614

If McGuinness ended up vying for the presidency with Mitchell (in this case replicating the above analysis involving McGuinness and Higgins, but swapping the positions of Mitchell and Higgins) he would need to be almost seven percent ahead of what would be his main challenger on the first count in order to win  – McGuinness 23.8%, Norris 19.5%, Higgins 12.1%, Davis 12.1%, Mitchell 16.7%, Gallagher 10.2%, Scallon 5.6%:

	1st	2nd	3rd	4th	5th	6th
Norris	297,220	305,712	313,709	325,414	360,103
Higgins	183,993	201,827	249,806	315,357
McGuinness	361,830	373,719	399,308	406,331	434,713	481,527
Davis	183,993	206,922	234,110
Mitchell	254,760	266,649	293,837	343,001	424,993	481,506
Gallagher	155,687	159,933
Scallon	84,920

So, in order to be in a position to win, both Norris and McGuinness will either need to gain up to another ten percent in terms of first preference votes over the rest of the campaign (leaving them with a percentage share of the first preference vote in the high 20s, at least), or else significantly improve their transfer attractiveness, or else hope that they are vying at the end with a less transfer-attractive candidate, such as Mitchell. Both are helped somewhat by the large number of candidates in the field, which means also that there will be probably a significant increase in numbers of non-transferable votes as supporters of lower placed candidates fail to give preferences to all candidates.

However, unless either Norris or McGuinness significantly improve on their first preference tallies or become more transfer-attractive, it is hard to see either winning in a head-to-head battle for the presidency with Higgins. At present, and being mindful of these transfer patterns, the main threat to a Higgins victory would appear to be a surge in support for Mary Davis (or for Gay Mitchell, to a less evident extent) as the campaign progresses. If the first poll of the official campaign shows momentum building for Davis, she could well offer a serious challenge to Higgins for the ultimate prize.

****************************************************************************************

In terms of the poll measuring current levels of party support, support levels for both government parties are seen to be roughly 3% lower than the levels achieved in the February 2011 election, but this will be a significantly most disappointing poll finding for Fine Gael given that party’s very strong showings in a range of opinion polls in the months since the election. The post-election honeymoon may finally be over for Fine Gael – previous polls suggest that this ended somewhat earlier for Labour and this poll figure compares relatively favourably with some other polls held during the post-election period but most notably the September 4th Sunday Times/Behaviour & Attitudes poll. Offering even better news for Labour is the finding that Michael D. Higgins still stands poised to win the presidential election – Higgins’ significant ability to attract to transfers would see him win the election based on these poll
standings despite David Norris currently leading the field in terms of first preference votes.

The September 25th Red C-Sunday Business Post poll puts national support levels for the main political parties and groupings as follows: Fine Gael 33% (down 8% relative to the last such poll in May 2011), Labour 16% (down 3%), Fianna Fail 15% (down 1%), Sinn Fein 15% (up 4%), Independents and Others 21% (down 8%). The significant trends here relative to the previous poll in May relates to the significant drop in support for Fine Gael, with support levels for the government parties falling by a combined level of 11%.

Based solely on assigning seats on the basis of the constituency support estimates (simply using a d’Hondt method to determine which party wins the seats), party seat levels would be estimated as follows: Fine Gael 69, Labour 28, Fianna Fail 18, Sinn Fein 20, Others 31. When the factors of vote transfers and vote splitting/management (based on vote transfer/management patterns oberved in the February 2011 election) are accounted for and constituency marginality levels at the February 2011 election taken account of, the party seat levels would more than likely be as follows: Fine Gael 70, Labour 32, Fianna Fail 16, Sinn Fein 21, Others 27.

The constituency support estimates based on the poll figures are as follows:

	FF	FG	LB	SF	OTH
Carlow-Kilkenny	25%	37%	14%	15%	9%
Cavan-Monaghan	16%	33%	4%	36%	10%
Clare	20%	41%	13%	0%	27%
Cork East	15%	35%	27%	17%	6%
Cork North Central	13%	23%	21%	22%	20%
Cork North West	23%	47%	12%	12%	6%
Cork South Central	25%	33%	16%	13%	13%
Cork South West	21%	47%	12%	12%	8%
Donegal North East	14%	27%	8%	34%	18%
Donegal South West	17%	16%	4%	43%	21%
Dublin Central	12%	18%	23%	19%	28%
Dublin Mid West	10%	28%	25%	18%	18%
Dublin North	14%	30%	22%	0%	34%
Dublin North Central	11%	35%	19%	8%	27%
Dublin North East	10%	27%	29%	18%	16%
Dublin North West	10%	15%	35%	32%	8%
Dublin South	8%	33%	15%	4%	41%
Dublin South Central	8%	21%	29%	20%	22%
Dublin South East	10%	33%	21%	6%	30%
Dublin South West	9%	25%	30%	26%	10%
Dublin West	15%	25%	24%	9%	26%
Dun Laoghaire	14%	34%	27%	0%	26%
Galway East	16%	40%	11%	9%	25%
Galway West	18%	27%	10%	9%	36%
Kerry North-West Limerick	10%	36%	16%	30%	9%
Kerry South	11%	29%	9%	0%	51%
Kildare North	13%	31%	25%	9%	22%
Kildare South	20%	32%	24%	10%	14%
Laois-Offaly	23%	30%	6%	16%	25%
Limerick City	19%	41%	17%	14%	8%
Limerick	19%	48%	16%	0%	17%
Longford-Westmeath	18%	37%	23%	12%	11%
Louth	13%	27%	15%	31%	14%
Mayo	14%	62%	4%	10%	10%
Meath East	17%	39%	18%	14%	12%
Meath West	15%	42%	11%	26%	6%
Roscommon-South Leitrim	12%	34%	7%	14%	32%
Sligo-North Leitrim	18%	33%	8%	20%	21%
Tipperary North	14%	21%	16%	9%	40%
Tipperary South	11%	30%	9%	7%	44%
Waterford	12%	34%	15%	15%	23%
Wexford	16%	32%	17%	9%	26%
Wicklow	9%	35%	14%	15%	27%

Based solely on assigning seats on the basis of the constituency support estimates (simply using a d’Hondt method to determine which party wins the seats), party seat levels would be estimated as follows:

	FF	FG	LB	SF	OTH
Carlow-Kilkenny	1	2	1	1
Cavan-Monaghan	1	2		2
Clare	1	2			1
Cork East		2	1	1
Cork North Central		1	1	1	1
Cork North West	1	2
Cork South Central	1	2	1		1
Cork South West	1	2
Donegal North East		1		1	1
Donegal South West	1			1	1
Dublin Central		1	1	1	1
Dublin Mid West		1	1	1	1
Dublin North		1	1		2
Dublin North Central		1	1		1
Dublin North East		1	1	1
Dublin North West			2	1
Dublin South		2	1		2
Dublin South Central		1	2	1	1
Dublin South East		2	1		1
Dublin South West		1	2	1
Dublin West	1	1	1		1
Dun Laoghaire		2	1		1
Galway East	1	2			1
Galway West	1	2			2
Kerry North-West Limerick		2		1
Kerry South		1			2
Kildare North		2	1		1
Kildare South	1	1	1
Laois-Offaly	1	2		1	1
Limerick City	1	2	1
Limerick	1	2
Longford-Westmeath	1	2	1
Louth		2	1	2
Mayo	1	4
Meath East		2	1
Meath West		2		1
Roscommon-South Leitrim		2			1
Sligo-North Leitrim	1	1		1
Tipperary North		1			2
Tipperary South		1			2
Waterford		2	1		1
Wexford	1	2	1		1
Wicklow		2	1	1	1
STATE	18	69	28	20	31

When the model is amended 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:

	FF	FG	LB	SF	OTH
Carlow-Kilkenny	1	2	1	1
Cavan-Monaghan	1	2		2
Clare	1	2			1
Cork East		2	1	1
Cork North Central		1	1	1	1
Cork North West	1	2
Cork South Central	1	2	1	1
Cork South West		2	1
Donegal North East	1	1		1
Donegal South West		1		1	1
Dublin Central		1	1	1	1
Dublin Mid West		2	1	1
Dublin North		2	1		1
Dublin North Central		1	1		1
Dublin North East		1	1	1
Dublin North West			2	1
Dublin South		2	1		2
Dublin South Central		1	2	1	1
Dublin South East		2	1		1
Dublin South West		1	2	1
Dublin West	1	1	1		1
Dun Laoghaire		2	1		1
Galway East	1	2			1
Galway West	1	1	1		2
Kerry North-West Limerick		1	1	1
Kerry South		1			2
Kildare North		2	1		1
Kildare South	1	1	1
Laois-Offaly	1	2		1	1
Limerick City	1	2	1
Limerick	1	2
Longford-Westmeath	1	2	1
Louth		2	1	2
Mayo	1	3			1
Meath East		2	1
Meath West		2		1
Roscommon-South Leitrim		2			1
Sligo-North Leitrim		2		1
Tipperary North		1	1		1
Tipperary South		1			2
Waterford		2	1		1
Wexford	1	2	1		1
Wicklow		2	1	1	1
STATE	16	70	32	21	27

Given that there are a number of other potential gains seen to be within Sinn Fein’s grasp (such as Roscommon-South Leitrim and Waterford), in addition to the seven seats gains suggested by the results of this analysis, this has to be seen as a good result for that party and suggests that the party may be making political gains resulting from their decision to contest the presidential election and to run a high profile candidate in Martin McGuinness. This is the first poll to show the combined support levels of the two government parties falling below the 50% level for the first time since the general election, but although both parties would be predicted to lose six seats each they would still be winning a more than sufficient number of seats to maintain a comfortable majority in Dail Eireann based on these figures. Fianna Fail will be relieved that these figures are decidedly better than those of last week’s poll in The Sunday Independent but on the other hand these poll figures suggest that the party is now being outstripped by Sinn Fein as the main party amongst the opposition groups. With a potential combined seat tally of 70 across Labour, Sinn Fein and left-leaning independents and smaller party (including ULA) candidates, the poll findings also suggest the potential of an alternative Left coalition government option, should Fine Gael and Fianna Fail over time move towards a political marriage.