Adrian Kavanagh, 21st September 2016
The 18th September Sunday Times-Behaviour & Attitudes opinion poll shows Fianna Fail still holding the position of the most popular political party in the state, standing some five percentage points ahead of Fine Gael. However, the two largest parties have lost some ground in this, the latest in the series of Sunday Times-Behaviour & Attitudes opinion polls, with Sinn Fein emerging as the main beneficiaries of this loss in support. Labour continues to regain the ground lost by the party in the opinion polls carried out in the weeks/months immediately following the February election. The 18th September Sunday Times-Behaviour & Attitudes opinion poll estimated party support levels as follows: Fianna Fail 28% (down 2%), Independents and Others 24% (NC) – including Social Democrats 1%, Anti-Austerity Alliance-People Before Profit 4%, Renua <1%, Green Party 2%, Workers Party 1%, Independent Alliance 4%, Other Independents 12% – Fine Gael 23% (down 2%), Sinn Fein 18% (up 4%), Labour Party 7% (up 1%). My constituency-level analysis of these poll figures estimates that party seat levels, should such national support trends be replicated in an actual general election, would be as follows: Fianna Fail 54, Fine Gael 43, Sinn Fein 27, Anti Austerity Alliance-People Before Profit 5, Labour Party 7, Social Democrats 1, Green Party 0, Independents 21.
Independents and Others: Support levels for the Independents and Others grouping remain at a relatively high level, standing roughly equivalent to the level that they stood at in the weeks leading up to the February 26th election and at the contest itself. In the lead up to the February election, seat levels for the Independents and Others grouping were notably harder to glean than was the case for the larger political parties, given that this is a very large and diverse grouping. As the general election showed, in some cases increased support levels for a smaller party were simply absorbed by the increased number of candidates within that party/grouping, meaning that increased support levels did not translate into increased seat numbers – as happened in relation to the 2002 and 2007 Sinn Fein votes and seats levels, as well as the support/seat level figures for the Social Democrats and Renua in 2016. It was impossible to apply the model in the case of the smaller parties/groupings ahead of the 2016 General Election, as many of these smaller parties and groupings did not exist in 2011 or had contested only a relatively small number of constituencies at the 2011 General Election. But, with a good number of these parties having contested a number of constituencies in the 2016 election, it is now possible to treat them as separate entities in this model, although figures can not be gleaned for constituencies that these parties did not contest on February 26th, as in the case of the Social Democrats and the majority of Dail constituencies).
The nature of the Independents and Others grouping means that support levels do not usually translate as neatly into seat gains as would be the case with the larger parties such as Fianna Fáil, Sinn Fein and Fine Gael. Vote transfer levels across this grouping will generally not prove to be as strong as the extent of intra-party vote transfer levels enjoyed by the larger political parties, who in turn often enjoy a “seat bonus” at most general election contests. Against that, candidates from the Independents and Others grouping may be more likely to attract transfers from political parties than would be the case with candidates from other political parties, as indeed proved the case in certain circumstances (most notably Maureen O’Sullivan in Dublin Central) at the 2016 election. And, of course, the better that candidates from this grouping fare in the different constituencies, the more likely they are to attract vote transfers given that they will be placed higher up in the order of first preference vote levels and hence well placed to last longer/until the end of the election count.
The problem for Labour: Like the canary in the coal mine, the constituency level poll analyses consistently warned in the years leading up to the 2016 General Election that low levels of Labour Party support nationally would translate into very low seat levels for that party. Given the Labour Party’s geography of support, but also given the increased level of opposition the party faces on the left of the political spectrum (from Sinn Fein, Anti Austerity Alliance-People Before Profit, the Social Democrats and other left-wing groupings or left-leaning independent candidates). These analyses consistently argued that Labour would struggle to convert votes into seats if their national support levels fall below the 10% level, as indeed proved to be the case with the February 26th election. If the party’s support levels nationally were to fall even lower than the 6.6% level won in the general election, then the party would be losing even more seats – especially given the narrow margins that Labour candidates won seats by (such as Willie Penrose in Longford-Westmeath) in that election. The 7% support level in this poll puts Labour slightly ahead of the level of support won by that party at the February 2016 contest. On the 7% support level in this poll, Labour would need a lot of luck to end up winning ten seats, or more, at a general election contest – on the figures in this analysis the party would really only be competitive in a small number of constituencies, with Wexford and Cork East emerging as the most notable examples here. However, they could edge out another seat or two, if problems exist with the vote management approaches of other political parties, as indeed happened at the February 2016 contest.
How this model works: Constituency support estimates for different parties and groupings form the basis of the general approach taken with this analysis. This seeks to ask the following question in relation to different opinion poll results – what do these poll figures mean in terms of the likely number of Dail seats that could be won by the different parties and groupings on those national support levels? Although the Irish electoral system is classified as a proportional electoral system, the proportion of seats won by parties will not measure up exactly to their actual share of the first preference votes, mainly because geography has a very significant impact here. First preference votes need to be filtered through the system of Irish electoral constituencies (and the different numbers of seats that are apportioned to these). In order to address this question, I estimate what the party first preference votes would be in the different constituencies, assuming similar (proportional) changes in party vote shares in all constituencies to those that are being suggested by a particular opinion poll. This of course is a very rough model and it cannot take appropriate account of the fact that changing support levels between elections tend to vary geographically, while it also fails to take account of the local particularities of the different regions in cases where no regional figures are produced in association with different national opinion polls meaning that there is no scope to carry out separate regional analyses based on these poll figures.
Thus constituency support estimates for different parties/groupings will be over-estimated in some constituencies and under-estimated in others, but the expectation would be that the overall national seat figures figures estimated will be relatively close to the true level, given that over-estimates in certain constituencies will be offset by under-estimates in others. Based on these estimated constituency support figures, I proceed to estimate the destination of seats in the different constituencies. The constituency level analysis involves the assigning seat levels to different parties and political groupings on the basis of constituency support estimates and 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 observed in the February 2016 election). Due to unusually high/low support levels for some parties or political groupings in certain constituencies in the previous election, the model may throw up occasional constituency predictions that are unlikely to pan out in a “real election”, but of course the estimates here cannot be seen as highly accurate estimates of support levels at the constituency level as in a “real election” party support changes will vary significantly across constituency given uneven geographical shifts in support levels.
The model has been amended to take account of instances where general election candidates have changed their political party/political grouping, as with the case of Stephen Donnelly’s (Wicklow) defection from Social Democrats and return to the Independent ranks. In this instance, part of that candidates February 2016 support level is transferred from the old party’s tally to the new party/political grouping when carrying out analyses with this model.
The point to remember here is that the ultimate aim of this model is to get an overall, national-level, estimate of seat numbers and these are based, as noted earlier, on the proviso that an over-prediction in one constituency may be offset by an under-prediction in another constituency. This model does not aim, or expect, to produce 100% accurate party support and seat level predictions for each of the 40 constituencies. These analyses simply estimate what party seat levels would be, should such national support trends be replicated in an actual general election. For a variety of reasons (including the impact of high levels of undecided voters in a specific poll), the actual result of an election contest may vary from the figures suggested by an opinion poll, even if the poll is carried out relatively close to election day, or on election day itself as in the case of exit polls, but the likelihood of such variation is not something that can be factored into this model. Vote transfer patterns of course cannot be accounted for in the constituency support estimate figures, but I do try to control for these somewhat in my set of amended seat allocations.
Constituency Support Estimates: The constituency support estimates based on the Sunday Times-Behaviour & Attitudes poll (18th September 2016) are as follows:
|Cork North Central||30%||15%||7%||24%||6%||18%|
|Cork North West||40%||29%||0%||9%||19%||4%|
|Cork South Central||45%||22%||4%||15%||6%||6%|
|Cork South West||22%||29%||7%||11%||29%||1%|
|Dublin Mid West||18%||23%||5%||29%||11%||14%|
|Dublin Bay North||18%||18%||8%||16%||25%||15%|
|Dublin North West||16%||12%||9%||39%||6%||18%|
|Dublin South Central||14%||13%||8%||30%||19%||16%|
|Dublin Bay South||15%||31%||13%||14%||5%||23%|
|Dublin South West||16%||20%||7%||18%||17%||22%|
Note: “OTH” here refers to the total support/seat levels estimated for the smaller parties – as the published version of an earlier post showed, it gets “messy” if there are too many columns in the tables here!
Seat Estimates: 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:
|Cork North Central||2||1||0||1||0||0|
|Cork North West||2||1||0||0||0||0|
|Cork South Central||2||1||0||1||0||0|
|Cork South West||1||1||0||0||1||0|
|Dublin Mid West||1||1||0||2||0||0|
|Dublin Bay North||1||1||0||1||2||0|
|Dublin North West||1||0||0||2||0||0|
|Dublin South Central||1||0||0||2||1||0|
|Dublin Bay South||1||2||0||1||0||0|
|Dublin South West||1||1||0||1||1||1|
Amended Seat Estimates: The seat estimates also need to take account of the candidate and competition trends unique to the different constituency. Here, the model is amended to account for seats that may be won or lost on the basis of estimates here being based on support levels derived due to a large/small number of candidates contesting the election. For now – the vote transfer/management patterns evidenced in the February 26th contest – are also being factored in, although the assumption that candidate numbers would remain the same as in the February contest is relaxed, given the not unrealistic assumption that parties, such as Fianna Fail, will run more candidates in constituencies where their candidates were elected with very large surpluses on the First Count. Taking these concerns into account, the amended seat allocations across the constituencies would look more like this:
|Cork North Central||1||1||0||1||0||1|
|Cork North West||2||1||0||0||0||0|
|Cork South Central||2||1||0||1||0||0|
|Cork South West||1||1||0||0||1||0|
|Dublin Mid West||1||1||0||1||0||1|
|Dublin Bay North||1||1||0||1||1||1|
|Dublin North West||1||0||0||2||0||0|
|Dublin South Central||1||1||0||1||1||0|
|Dublin Bay South||1||2||1||0||0||0|
|Dublin South West||1||1||0||1||1||1|
Potential Governments?: In some ways the political landscape that might emerge out of an election, where party/grouping support levels were to be the same as in these opinion poll figures, would be very similar to the landscape that emerged after February 26th, but for the fact that the positions of Fianna Fail and Fine Gael would be reversed here. Based on these seat estimates, a Fine Gael-Fianna Fail “grand coalition” would a two-party coalition that would be capable of forming a government without needing the support of others (97 seats in total). A potential coalition/alliance involving Fianna Fail and Sinn Fein (81 seats in total) would also exceed the magic 79-seat number. Given the changed context set by the new government arrangement, a minority Fianna Fail administration, involving independents and smaller parties potentially, could be a very likely outcome in this particular scenario.
Why are Labour likely to win less seats than in 1987 on a low national support level?: The seat level estimates in all of the recent poll analyses for the Labour Party have been stark and the estimate in this poll is especially stark (highlighting the fact that the PR-STV system is proportional, but only to a limited extent). Previous analyses have, moreover, suggested that, especially given the increased competition on the Left from Sinn Fein, other smaller left of centre parties and left-leaning independents, that it will be a struggle for Labour to win seats in most, if not all, constituencies if the party’s national support levels fall below the ten percent level, as has been shown in similar analyses of most recent polls. The further the party falls below this ten percent level, the more problems Labour faces in terms of winning seats. Labour would be in serious trouble if their national support levels fall below ten percent as the party is also facing a “perfect storm” from electoral geography and changed competition levels. These factors include the reduction in Dail seat numbers (from 166 to 158) and other changes made to general election boundaries by the 2012 Constituency Commission (which militated against Labour while seeming to advantage other parties, but notably Fianna Fail) as well as the increased competition the party now faces on the Left from Sinn Fein, other smaller left-wing parties and left-of-centre independents, as well as from Fianna Fail. When Labour support levels fell to similarly low levels in the late 1990s and early-to-mid 2000s, the party was in a position to be helped (as in the 1997, 2002 and 2007 General Elections) by transfers from lower placed candidates from the smaller left-wing parties. But on these constituency-estimate figures outlined in these analyses Labour Party candidates would find themselves polling below candidates from Sinn Fein, the Socialist Party, the Workers and Unemployed Action Group or the People Before Profit Alliance, or left-leaning independents, in a number of constituencies. Instead of being in a position to possibly benefit from vote transfers (which themselves would be likely to dry up in any case), the Labour candidates would now in a number of cases be eliminated before the final count and would be providing the transfers to see candidates from other left-of-centre political groupings over the line. (If we look at the 1987 case study – we see Labour won 6.5% of the vote in the 1987 General Election and won 12 seats, but it is also worth noting that they did not contest nine constituencies in that election, whereas their 7% national vote is being distributed across all forty constituencies in this analysis, as with the most recent general elections in which Labour has contested all constituencies. In two of the twelve constituencies in 1987 where Labour won seats – Dublin South-Central, Dublin South-West, Galway West and Wexford – vote transfers were crucial in ensuring Labour won these these seats – i.e. Labour candidates were outside the seat positions on the first count but overtook candidates with higher first preference votes as counts progressed due to transfers from other candidates.
|Constituency||FPV||Total Poll||Quota||% FPV||Lab/quota|
Voting statistics for constituencies in which Labour won seats at the 1987 General Election. The table above shows that there was no constituency in 1987 in which a Labour candidate exceeded the quota and indeed successful Labour candidates, Ruairi Quinn and Michael D. Higgins won seats in their constituencies despite winning less than half of the quota in their first preference votes. In addition, Dick Spring came within a handful of votes of losing his seat in Kerry North.)
The Labour Party has tended to fall below the ten percent level in most opinion polls over the past few years, as in the case of this Behaviour & Attitudes opinion poll. Labour seat level estimates in most of the poll analyses I have carried out in the years leading up to the recent general election were quite stark, highlighting the fact that our PR-STV electoral system is proportional but only to a limited extent. The party’s seat levels in the actual election generally tended to support the analysis offered by this model. The further Labour support nationally falls below the ten percent level, the more difficulties it will face in terms of winning seats. This proved to be the case in February 2016 as Labour was left to face a “perfect storm” from the combined effects of boundary changes, electoral geography and changing political competition patterns. These factors explain why Labour faces greater challenges in translating lower levels of national support into seat numbers than it did back in 1987 when the party won 12 seats with just over six percent of the national vote.
- The size of the Dáil was reduced from 166 seats to 158 at the recent general election, but my analysis of the effects of these boundary changes suggested that Labour will be more adversely effected by these than other parties, such as Fianna Fáil in particular, would be. Had the new boundaries been in place in 2011, it is estimated that Labour would probably have won three or four fewer seats, while Fianna Fáil may have won two or three more seats!
- While there is a distinct geography to Labour Party support levels over and above the more “catch-all” trends traditionally associated with Fianna Fáil and (to a lesser degree) Fine Gael, there is not the same concentration of support into a small number of constituencies that one has observed in past contests with smaller parties such as the Green Party (especially in the 2002 and 2007 contests) and potentially parties such Renua Ireland, the Social Democrats and the Anti Austerity Alliance-People Before Profit in the upcoming election. In the latter cases, lower support levels nationally often translate into much larger support levels in a small number of stronger constituencies, allowing these to pick up a number of Dáil seats. In the case of Labour, the same spiking of support in a small number of stronger constituencies is not evident. As evidenced at the 2016 contest, a small share of the vote nationally (especially if Labour contests all, or most of, the Dáil constituencies) would translate into a sufficient level of support to allow them to challenge for seats in only a very small number of their stronger constituencies. (Labour were also not be helped by the level of defections and retirements amongst its cohort of TDs, especially given that these involve many of the party’s stronger constituencies at the 2011 contest. This was perhaps most evident in the party’s failure to win seats in some of its strongest constituencies in 2011, namely Dublin South-West, Dublin South-Central and Dublin North-West.)
- In 1987, Labour won 12 seats even though the party never exceeded the quota in any of the constituencies being contested by the party in that election (and it is worth remembering that Labour did not contest every constituency in 1987). Indeed, Labour won nearly half of their seats in constituencies where they had won little more than half a quota in terms of first preference votes – or even less than half a quota in the case of the Galway West and Dublin South-East constituencies. Labour were helped in this instance as their candidates were in a position to pick up vote transfers from lower-placed left-wing candidates, as well as lower-placed Fine Gael candidates (arising from that party’s drop in support in 1987 and also some instances of poor vote management). In 2002 and 2007 Labour were also able to translate their national support levels into a higher proportion of Dáil seat levels due to Labour candidates being ahead of candidates from other left-wing groupings and hence in a position to win transfers from these. In the context of low Labour support levels nationally in the 2016 election, however, the trend in a number of constituencies was instead for Labour Party candidates to be polling below candidates from Sinn Fein, the Social Democrats or the Anti-Austerity Alliance-People Before Profit, or left-leaning independents. Instead of being in a position to possibly benefit from left-wing vote transfers (which themselves were weaker in any case, as was predicted earlier in data provided in a Sunday Independent-Millward Brown poll in the summer of 2015), in this context Labour candidates were eliminated before the final count in many constituencies and potentially providing the transfers to ensure the election of candidates from other left-of-centre political groupings (or Fine Gael candidates). In previous elections, Labour might still have been able to translate lower support levels into seats in constituencies such as Dublin Bay-South, Dublin Bay-North, Clare and Dublin Rathdown, but this proved not to be the case in the 2016 contest. Ironically, with the party now out of government, they may find that their ability to win vote transfers improves again over time, along the lines of the trends that were observed for the Green Party at the recent general election.