Will the SDLP’s change in MLAs yield electoral success in May?

There has been an interesting changing of the guard within the SDLP over the last year across all levels of the party.

We all know that the leadership team was given a dramatic shake up in November when Colum Eastwood ousted Alasdair McDonnell as party leader and Fearghal McKinney took the deputy post from Dolores Kelly.

But another shift has been happening that is noteworthy and that is the shake-up in the party’s assembly group.

Over the past year we have seen younger MLAs co-opted into the Assembly group with Claire Hanna (35), Daniel McCrossan (27), Gerard Diver (51) who are all significantly younger than the MLAs they are replacing.

These changes have brought the average age of an SDLP MLA from 53 to just under 50 years old.

On top of this we have the departure of other party stalwarts like Dominic Bradley, John Dallat and Alban Maginness.

Bradley will either be replaced by current MLA, Karen McKevitt or Justin McNulty, Maginness is likely to be replaced by Nichola Mallon and there are number of names in for Dallat’s seat.

Should the party hold all of their current seats that would mean that roughly half (7 when you include McKinney) of the SDLP’s group would be freshly elected at the next assembly election.

My question for you lot is, will this punt work out? I know youth is no guarantee of political success, but can this reshuffle and attempt at renewal yield some success for the party on May 5th?



  • mjh

    In the short-term, which means the next election, the answer is “No”. On average around 1 in every 5 first preference votes is cast for the candidate rather than the party. (That is the proportion of voters who do not give their second preference to a candidate from the same party when one is available.)

    If the candidate is a long-serving, popular elected representative this personal vote can rise to 1 in 4, or even (as in the case of Margaret Ritchie) to 1 in 3.

    So replacing old warhorses with bright young things is likely to suppress a party’s vote the first time round. Of course there could be a dividend at future elections if the new member can build a bigger personal following than the old.

  • Ernekid

    As I’ve said before the SDLP need to figure out what their USP will be for the next 5 years from 2016-2021 and why exactly people should vote for them instead of voting for Sinn Fein or not voting at all. Sinn Fein have a fairly strong brand as they’ve established themselves as a mainstream Irish nationalist party that has a presence in all 32 counties of Ireland and will have an increased presence in Dáil Éireann following the Irish General Election. Sinn Fein can easily frame itself as the party that protects and promotes Irish nationalist interests and culture in local government and act as a check against the Shitehawkery of unionism. The SDLPs position is much less clear. It can’t run as the party of peace as the Peace is nearly 20 years old. They don’t seem to be very progressive as marked by their lukewarm approach to LGBT rights and women’s issues. They haven’t set out a Big idea that’ll mark them out from their political rivals.

    Why should I vote for you? Is a very good question to ask the SDLP. I’m not sure if we’ve gotten a good answer yet.

  • Kevin Breslin

    I don’t think the replacements mentioned will do much damage, McCrossan is a very active candidate, Justin McNulty probably has a bigger reputation than Dominic Bradley and McKevitt’s already established herself, Gerard Diver is experienced enough to have previously contested an Assembly contest before, Fearghal McKinney and Claire Hanna are well established too.

    After the elections the change could mean after perhaps Dolores Kelly, Patsy McGlone and Alex Attwood, Colum Eastwood would be joint 4th longest serving SDLP MLA with Mark H Durkan and maybe Karen McKevitt. Sean Rodgers comes next after replacing Margaret Ritchie after she stood down from her seat.

    Interesting to see Brenda Stevenson putting herself forward in Dallat’s constituency. I wonder if she has moved.

  • Robin Keogh

    Such is the nature of Irish Political Culture it is quite clear that the most effective way of getting your face into the minds of the electorate is sheer hard graft on the ground. Yes, the media’s presentation of a party or individual can have an impact but I think the growth of modern SF especially in the south shows that even a vindictive, determined and disengenous medai and political establishment cant beat the power of the personal touch. Showing up to residents meetings or attending local funerals wont do the trick. Popping up at public events and ceremonies might get you known but the real prize awaits those who genuinely show a sincere interest in people and the everyday problems they face trying to get from dawn to dusk with work, family other commitments. Whatever age, wannabe representatives need to walk and talk at least in equal measure and be known in their communities for the hard work they do and not simply by who they are or what party they represent.

  • mjh

    In South Down, Newry & Armagh and East Londonderry these replacements would not reduce the number of SDLP seats, although it is likely to reduce the SDLP lead over SF in South Down, probably quite substantially.

    In South Belfast the risks have been reduced by getting Claire Hanna into place sooner rather than later. Plus she brings a substantial personal vote of her own from the Balmoral part of the constituency ( at least 1 in 4 there). In an already challenging contest however there is real risk that the absence of McDonnell from the ballot paper, with the personal vote he had built throughout the constituency over years, could help to cost the party the second seat. Whether this will be offset by the loss to Alliance of Anna Lo’s personal vote remains to be seen but the recent background trend in the party vote has been for Alliance to be on the up and the SDLP down.

    The biggest risk from candidate substitution could well be North Belfast, where Council and Westminster votes suggest that the SDLP Assembly seat is already on a knife’s edge. We don’t have any recent data on Alban’s personal vote – but after 30 years it may be bigger than Nicola Mallon has yet had time to build.

  • Kevin Breslin

    I’m not sure if there’s going to be any reduction in Newry & Armagh, it is a strong ticket. McNulty increased on Bradley’s Westminster vote. South Down does seem to have a couple of established Councillors with an MLA. I don’t know who is standing in the other one but the seat does seem safe and John Dallet has been positive about young replacements. Foyle was retained with the MP gone from the ticket, I’d imagine that the SDLP may improve their first count overall at Assembly level with two candidates increasing their profile.

    Alliance having any weakness in South Belfast would be a mixed blessing, because it was Alliance transfers that brought SDLP 2nd returned candidate Conall McDevitt through against Ruth Paterson, I think Paula Bradshaw and Duncan Morrow collectively aren’t going to lose too much of Lo’s vote if any. Would be interesting to see how many Alliance transfers would go to Pengelly and her running mate, which I still believe will be Stalford.

  • Skibo

    Are there extra seats for SDLP in north and south Antrim where the nationalist vote seems to be 5 % down from where it should be?
    I wonder just how safe Dolores Kelly will be if there is a strong push by SF for the second nationalist seat?

  • Granni Trixie

    I would have thought that taking McDonnel off the ballot paper is to the advantage of the SDLP, bearing in mind his popularity is in decline for obviousreasons.

  • Granni Trixie

    What this post brings to the fore is the necessity for (well run?) parties to do some legacy planning so that they are not caught on the hop. This is done behind the scenes if at all so it’s difficult to accurately read if changes in personnel in the SDLP (or any party) is the outcome of planning or expediency.

  • Granni Trixie

    I think it will be fascnating to observe how Pengelly does given she is the candidate who is there because her leader favoured her and she has no record of work on the ground,indeed a DUP member resigned in protest at her taking on role of MLA. For these reasons I would expect even diehard DUP voters will hesitate before voting for her. It will be educational to see what happens in view of these factors.

  • Colin Lamont

    I’m not sure about that. Mcdonnel has been the most established and high profile rep in that area for a long time. He gets plenty of personal and tactical votes; if he hadn’t been the Westminster candidate I believe the SDLP vote would have been even worse than it seemed at the time. Consequently the South Belfast seat would have been lost.
    More broadly, although the SDLP on a really bad day could lose up to 3 Belfast MLA’s, I think it more likely that there will be no change especially considering their transfer friendliness.

  • Granni Trixie

    I live in SB and I assure you that
    McDonald personal votes are on the decline ,

  • mjh

    In order of statistical probability, in all cases taking the 2011 Assembly election as the base:

    Upper Bann: If there were a general swing from SDLP to SF, Dolores Kelly’s seat would be one of the first to fall. Only 175 votes would have to move from SDLP to SF. That is a 0.4% swing (i.e. 0.4% of the first preference cast for all candidates of all parties in the constituency.)

    North Antrim: It would require a 0.8% swing from DUP to SDLP to deliver an SDLP gain. The same result would come from a 5% increase in the nationalist vote provided that all the increase went to the SDLP, or that any that went to SF transferred fully to SDLP. Since in real life that does not happen the nationalist increase would probably need to be around 6% or 7%. This of course assumes no change in the votes for other parties.

    South Antrim: Would need a 2.1% swing from DUP to SDLP. Alternatively, theoretically a 13% increase in the nationalist vote would do the same job, although again this would probably need to be 15% to 16% in real life.

  • Gopher

    The question the SDLP have not asked themselves is what happens if they win. They have positioned themselves that the constitutional issue like SF is their reason for existence, therefore people could end up with a thirty two county socialist republic if they enable an SDLP victory. The declining Nationalist vote and in particular the drift of new voters to People before Profit, Alliance and Greens, parties that have embranced the the logic of the GFA rather than have their vote interepted on a single issue seems to have validated this assertion.

  • Brendan Heading

    I’m not sure at all about McDonnell attracting tactical votes. A reading of the election results suggests the opposite.

    On paper (excluding personal aspects for the moment) there was a very real likelihood that the DUP could attract enough tactical votes to beat McDonnell. In the end, significant tactical voting in favour of the DUP did happen, although it fell just shy of what the DUP needed to win the seat. In this scenario, in a liberal constituency like SB where McDonnell would seem like the lesser of two evils, you would have assumed that people would have felt motivated to stop the DUP.

    In reality, no tactical shoring-up of McDonnell happened. Normally in a scenario like this Alliance get squished; yet Paula Bradshaw managed to beat Anna Lo’s previous record quite significantly, which tells us that every single Alliance voter, plus a few other people, felt no compulsion whatsoever to support McDonnell against the DUP, even though Anna herself was not the candidate. The fact that SF even stood their candidate shows that they were prepared to run the same risk.

    McDonnell’s vote fell by 16.5 percentage points and he very nearly lost the seat. It is pretty much an accident that he didn’t.

  • mjh

    Tactical votes and personal votes should be considered separately, although clearly it can at times be difficult to separate them out.

    In terms of tactical voting, I read the figures as suggesting that quite a lot of it went on in South Belfast in the Westminster election, and that the principal beneficiary was the SDLP. Indeed without it they would certainly have lost the seat.

    As you say the SDLP vote share was down dramatically from the previous Westminster election, but SF pulled out of that contest leaving the SDLP as de facto nationalist unity candidate. Nevertheless, against a general background of falling shares across NI at every election since, the SDLP share was approximately 6 percentage points higher than in the 2014 council elections, and even marginally higher than in the 2011 Assembly election. And this despite clearly having lost votes to SF which was about 1.5 percentage points up on its 2014 score.

    Yes Alliance was up 2% points on its previous Westminster result. But given the growth since then in centre party support in the constituency, particularly for Alliance, this still suggests a squeeze. The share for centre parties dropped by about 5 or 6% points from the 2014 elections. Alliance itself was down 2% points – although the absence of NI21 from the Westminster ballot should have put it about 2% points up.

    The UUP was down 1% point, and the DUP up about a half of one. It doesn’t look like there was significant tactical voting for the DUP.

  • Nicholas Whyte


    I largely agree on your “swing” statistics (though surely only 171 votes in Upper Bann), but I think you need to be clear that your “nationalist increase” figures are relative to the size of the nationalist vote, not to the total electorate (which is the basis for your “swing” stats). I’m also not sure that I agree with your figures on that point.

    In North Antrim, the SDLP lost to Jim Allister by 612 votes. Obviously if 307 Allister voters (0.8%, as you rightly say) had voted SDLP instead, Declan O’Loan would have kept his seat. If instead an extra 613 votes came out of nowhere, that’s 6% rather than 5% of the total Nationalist first prefs as cast (9834). But I think it is less confusing to say that the Nationalist share of the overall vote would need to increase from the actual 24.4% to 25.5% (9834+613 divided by 40313+613), with the gains entirely benefiting the SDLP, lifting them from 9.1% to 10.2%. It doesn’t seem an impossible stretch to me in a good year.

    In South Antrim you’ve adjusted for undistributed surpluses (as I would have done) making the losing margin more like 1400 votes than the actual 1077. I make that 17% of the Nationalist vote as cast, but again I’d prefer to express it as needing the Nationalist vote share to rise from 25.1% to 28.2%, with the gains entirely benefiting the SDLP, lifting them from 10.6% to 13.7%. That is a further stretch.

    You left out the closest result of the election, Fermanagh and South Tyrone, where the SDLP lost to SF by only 62 votes. A swing of less than 0.1% would win that back. Likewise in Newry and Armagh, the effective margin between the third SF seat and the second SDLP runner was only 130-140 votes, requiring a direct swing of 0.14%. Neither case requires a shift in the overall Nationalist vote share, though obviously a very small SDLP increase holding all else steady would be sufficient.

    And in SDLP / Unionist marginals, you left out Strangford, where a direct UUP swing of 0.8% to the SDLP would have got them another seat, or a rise in the Nationalist vote share from 11.5% to 12.9%, with the gains entirely benefiting the SDLP, lifting them from 8.5% to 9.9%.

    So on paper, the potential for fightback is there. There are four or five moderately winnable seats (not really sure about South Antrim) and a couple of longer shots that I haven’t mentioned (possible gain from Unionists in South Down, possible gain from Unionists or Alliance in Lagan Valley). But without an incumbent MLA, which is a problem in every case except Newry and Armagh, it’s much more difficult to mobilise resources effectively.

  • mjh

    Nicholas, thanks for your comments.

    I agree with your calculations on FST and Strangford. The reason I did not mention them was that they were not part of Skibo’s question.

    Yes the “increase in nationalist” vote figures are percentage changes to the total nationalist vote only, and it would have been better if I had made this absolutely clear.

    Yes, if one were examining the SDLP vote in isolation it would be clearer when discussing the effects of an improvement in turnout to express the changes solely in terms of the vote shares for that party. However, Skibo asked a specific question about the effects of an increase in the total nationalist turnout. It is a perfectly reasonable question but, unfortunately, a more “messy” one, and I have attempted to answer the question in the terms in which it was posed.

    As you say our calculations for the minimum increases in nationalist share required differ slightly. Mine are a little lower than yours because I have calculated them in a different way. Your figures take account of the fact that an increase in nationalist votes also increases the overall total of all votes cast. That would also increase the quota.

    So for example in North Antrim the 0.4% swing required to deliver a seat from TUV (I incorrectly said DUP) to SDLP can also be expressed as a swing of 0.06 of a quota. Adding 5% to the total nationalist vote gives a new nationalist total of 10,326 (9,834 + 492) and a new overall total of all candidates of 40,805 (40,313 + 492). The quota therefore rises from 5,760 to 5,830. The increased nationalist vote of 10.326 is 1.77 of the new quota. This is 0.06 higher than in the 2011 base figures. The unchanged total unionist vote was 4.97 of a quota in 2011 base, but drops 0.06 to 4.91 against the new higher quota. This gives us the quota swing of 0.06 required to shift the seat. (Assuming all the increase goes to the SDLP or transfers entirely from SF.)

    I doubt that anyone else is still reading this, but I would be interested to know if my approach is incorrect.

    By the way I’ve double checked Upper Bann and stand by 175. ((5787.07 minus 5437.50) divided by 2, rounded up to a whole number.)

  • Anyone else think Eastwood has been surprisingly, certainly exceptionally, quiet as a new leader. Absent at present.

  • Kevin Breslin

    Even his supporters must admit McDonnell having three jobs, (especially one that takes him out of Ireland) at once didn’t exactly help him put his best foot forward.

  • Kevin Breslin

    The tactical votes for Bell to beat McDonnell would’ve came at the expense of the UUP of any party, Brendan. Given the UUP vote halved, and an ex UUP candidate stood for the DUP here it is a reasonable assumption. Otherwise a second DUP at the expense of the UUP can’t be ruled out here.

  • Kevin Breslin

    The SDLP vote was down by less than 800, on 2005 levels, the last time Sinn Féin were in the contest. If the UUP didn’t contest this contest because of a pact, the party would be dead in South Belfast, there was no quid pro quo they could have offered. Since then South Belfast has more registered and more voting squeeze the SDLP vote share.

    Even if McDonnell was some how able to repeat 14,026 with or without Sinn Féin votes, the voter turnout increase would mean he’d be down 7% points anyway. repeating the 2005 tally of 10,339 on a turnout like this would put him down 13.5%. So effectively the net “800” lost votes since he first won the seat translates as contributing a 3 percentage point loss in my book.

    The 16.5% reduction I would attribute to in terms of ad hoc inference.

    7 percentage points – net new/returning voters not voting SDLP

    6.5 percentage points – votes swung to Sinn Féin/Alliance from 2010

    3 percentage points – lost votes

    Yes there is overlap and Yes who votes and who doesn’t vote with around 40% not voting in both, does give a very high margin for error.

    Realistically factoring the growth in turnout and electorate, the mathematically known certainty of people who must have voted in two elections (and even this assuming no migration/death of people between the periods from the original sample) would be less than 18% in reality.

    So if we are talking about individual’s personal choices changing between elections based on numbers alone you are probably closer to numerology than to empirically correct statistical hypothesis mechanism.

  • Gaygael

    Yeah. Totally concur!

    He did get stiffed though, in his first week by Robbo’s resignation and then the ‘fresh start’ deal, which prevented him gaining any media momentum.

    Silence since. Maybe there is a lot of internal stuff that needs done. I have a sense there is a digging in and taking whatever hits come their way in May and focus longer term. I think his camp linked the 5 losses document, to both undermine McDonnell and prepare s narrative that says any losses less than five is a success. He had not talked up any gains.

  • mjh

    Hello Kevin,
    Disappointed to see you using a whole paragraph of bold type, including the unpleasant charge of “numerology” in your reply above.

    If you are seeking to deny the conclusion that tactical voting is a factor in South Belfast you cannot do so by restricting your analysis to the 2005 and 2015 Westminster elections alone. You must also explicitly consider the Assembly and Council elections and account for the consistent pattern of higher SDLP shares at Westminster than at these other elections. (A pattern which is also present in South Down and Foyle.)

    If you are seeking to deny the conclusion that candidates can receive personal votes which have the potential to effect the outcome of an election you would need to consider the consistent evidence that certain candidates have, over years of successive elections, built up first preference votes which contain a considerably higher than normal number and proportion of voters who chose to give their second preference to a candidate from a different party.

    I believe it is important to try and understand the dynamics in the electorate and to discuss them. Otherwise we leave ourselves at the mercy of anyone to make unfounded claims, whether from lack of knowledge, wishful thinking – or more seriously as we saw in South Belfast at the last Westminster election – to cynically make misleading claims about party support in the attempt to influence how people vote.

  • barnshee


  • Kevin Breslin

    It was not meant to be offensive or a charge, but with 40% not voting in two elections, and the differing sample sizes. An exact statistical isn’t possible, ergo in my opinion even my own analysis may border on the numerology if we add supposition about voter behaviour. There is no real proof of voter behaviour from a sample size, there is very little evidence we can take here.

    Think about it, if the sample sizes were the same, same people with no a posteriori knowledge of who chose to vote between these elections, there is a possibility that everyone voted in one or other of the elections. So in theory, 80% could’ve voted in just one election and 20% in both. This cannot be ruled out by exact statistics. We know without making any other assumptions other than a consistent electorate that least 20% voted in both under assumptions of no to minimal change in the electorate.

    Bring in migration in and out of the constituency, births, deaths, even minor registration problems effecting the sample and the exact knowledge who who voted between the elections declines from 20%. By extension a full analysis of the trends without adding artificial a priori assumptions around

    However South Belfast does have a high level of “migrants” (not limited to internationals) in both directions, probably average to low in terms of births and deaths, there is a volatility in that

    Any analysis made from ANY two election results is subject to massive volatility. Your remarks about consistent evidence is merely speculative.

    It’s impossible to determine any exact evidence of who voted and who did not vote two 40% turnouts indicates anything from 60%-20% of the electorate voted in two elections added to the volatility from changes in the electorate and the uncertainty multiples. We don’t have any proof of who was and wasn’t voting. People didn’t put on paper who they voted for before and who they voted for after. You can only really compare one election by itself if you are talking about people.

    So I disagree, it is not important to understand the dynamics of the voters, we have a secret ballot where the system deliberately increases our level of certainty about the dynamics of individual voter intentions. There is really no way to tell.

    Effectively you cannot tell anything about the people for the most part, you can only speak about the population as a whole. We can talk about South Belfast as a whole, but we really can’t talk about the machinations of groups of voters.

    Voter shares can go down when a party support goes significantly up. I think it’s wrong to judge how well a party is doing on the basis of apathy of support in other parties, which is something another party really can’t or in my opinion should influence.

    A party can increase their vote share losing voters simply because there is greater apathy towards other parties, even parties pulling out of the contest, and decrease their vote share raising their vote because there is less apathy, because more parties enter. This is very relevant in Council and European Elections where fewer people vote.

    It would be difficult to envision a contest with a raised turnout and a larger number of parties not having a negative effect on a front runner.

  • mjh

    Firstly, Kevin, I fully accept that your posting did not intend any offence.

    Your posting raises a host of theoretical barriers to the possibility of analysing the results of elections for lessons on voter behaviour. But rather than testing their validity against the evidence, and then using them as useful tools to question and improve an analysis of the available data, you have concluded that no such analysis is possible. As a result you view any purported analysis as invalid.

    If we look at your main assumption they do not appear to hold up.

    You say that “it would be difficult to envision a contest with a raised turnout and a larger number of parties not having a negative effect on a front runner”.

    This assumption is immediately thrown into question by the three most recent results in South Belfast itself. In the Westminster contest the SDLP increased both its actual votes AND it’s percentage share relative to the 2011 Assembly elections. This despite a significantly higher turnout for Westminster – where 60.3% f the electorate voted compared to only 52.4% in the Westminster elections. Nine parties contested Westminster, and although two more stood in the Assembly this is not a significant factor, the two smallest parties receiving only 29 and 135 votes respectively.

    A comparison of Westminster and the Council elections the year before presents an even starker example. Only 6 parties contested all parts of the constituency, 6 others contested an average of half the constituency each ( of which the three smallest obtained only about 1.5% of the total vote between them). The turnout was approximately 50% to 51% – 10% points below the Westminster turnout. On your assumption the SDLP vote should have declined at Westminster – yet it went up from about 18%/19% to 24.5%.

    So the thing that was “difficult to envision” has actually occurred.

    Secondly, you state that “a party can increase their vote share losing voters”. Whilst I can see the mathematical truth in that proposition, you produce no evidence that it has actually happened in real life. Not even a suggestion of where or when this may have happened so that the assertion can be tested both for its validity and, if true, its impact. You may have that example – but without it this is just a bit of flack put up as a distraction.

    Finally, you appear to be suggesting that various differences in turnout and differential turnout; parties standing or not standing; increasing or decreasing rates of apathy afflicting the supporters of other parties; proportions of people registering, or moving in, or moving out or dying all make comparisons of performance between parties at one election, and by parties across a number of elections unfeasible. If this were so we would expect to see no patterns emerging, other than by pure chance. Indeed the performance of a party from one election to the next would not follow a trend, but would resemble a “random walk” in which the party share could go up or down irrespective of its actual core support or the tactical choices made or not made by those who would otherwise vote for other parties.

    A simplified model of this would present three options. Starting with its share in a Westminster election, its share in the closest Assembly election could be either 1) similar, 2) higher or 3) lower. There would thus be a one in three chance of the share being lower. The same would hold true for the nearest Council election.

    The chances of both being lower would be one in nine.
    The chances of all the Assembly and Council elections producing a lower share every time in all the comparisons shown on the ARK site from 1996 in South Belfast would be more than one in 2,000. The chances of this being true also in three constituencies would be poorer than 1 in 10 billion.
    Yet this is precisely what has happened in South Belfast, South Down and Foyle.

    Time, I think, either to accept that analysis might have some lessons to teach us; or to put up a convincing argument which confronts the contrary evidence in order to demonstrate that such analysis is unreliable.

  • Gaygael