An analysis of Transfers in the 2022 Northern Ireland Assembly Election
This is a long piece so I have tried to break it down into relevant categories. Some people hate looking at data so I have provided a summary of it here, others will want to see that so I have included some of it. Feel free to look for what you actually want to know and ignore the rest.
Intro.
In any STV election transfers will be important to the final result. In Northern Ireland we elect 90 MLAs, only 20 of them were elected on the first count meaning 70 MLAs were elected thanks to transfers either getting them over the quota or by keeping them ahead of their remaining rivals.
Each constituency elects 5 MLAs. The easiest way to be elected is to secure enough first preference votes to finish inside the top 5 on the first count but 6 candidates managed to win their seats despite finishing outside the top 5. Transfers were therefore vital to electing these 6 candidates.
The data presented below shows how voters distributed their preferences when voting, however it cannot tell the whole story. We can look for trends and get a good idea of how voters rank the parties and candidates in their mind but we have to remember that the transfers are dependant on who is left standing. For example, the SDLP had a total of 4,649.6 votes to transfer in Foyle, but they were transferred on stages 11 and 13 of the count. On stage 13 the only candidates left were the DUP and UUP representatives so the SDLP finished with 1,003.5 votes that were not transferred to anyone. Now, if we had the luxury of seeing the ballot papers it is likely there were actually many voters who gave preferences to the non-unionist candidates and stopped there, but these votes could not be transferred earlier because the SDLP needed them. These figures can never be perfect but can be useful for spotting which parties are likely to attract another party’s transfers.
Headlines
If you just want a few interesting points to take away from this then here they are:
The most transfer friendly party in the election was the DUP, winning 22% of all transfers and 21.2% of transfers when we exclude transfers from a party to itself. Interestingly when we exclude parties transferring to themselves the largest bloc is actually non-transferrable votes which account for 21.7%.
The DUP won almost 60% of all TUV transfers and the SDLP won around half of all Alliance transfers.
Sinn Féin voters seemed to be the most likely to express a preference for only one party.
Of the parties that stood in every constituency the least transfer friendly party was the TUV, attracting just 5.7% of all transfers.
After we remove parties transferring to themselves Sinn Féin has the largest drop, attracting just 0.2% more than the TUV. This shows that they are excellent at winning first preference votes but not good at attracting transfers.
Nationalists won significantly more transfers from the non aligned bloc than unionists and in some cases nationalists sent more transfers to the non aligned bloc rather than to another nationalist party.
Around 4% of all voters only voted for one candidate.
The rest of this piece will show all of this in more detail.
What data I have used and what was left out?
I have compiled data from every round of transfers, I have tracked how many votes transferred to which candidate by individual round and how many votes were non transferrable. I have then combed over this data and organised it to show which party’s votes have transferred to whom.
Unfortunately it is impossible to get a complete record as there are times when multiple candidates are eliminated at the same time, this makes it impossible to tell which votes transferred to which candidate. For example I have had to ignore one of the most interesting round of transfers in the whole election, the final stage in North Antrim. At stage 6 the 2 DUP candidates; Paul Frew and Mervyn Storey, were left standing alongside Patricia O’Lynn of Alliance. The quota was 8,438. Paul Frew was 1,878.8 votes short, Storey was 1,351.6 short and O’Lynn was 2,640.5 short. Matthew Armstrong (TUV) was eliminated on 2,726.3 votes and Eugene Reid (SDLP) was eliminated on 2,662 votes. Bizarrely both were eliminated together, given that the TUV surplus could have elected any of the candidates this stage should have been split in two. All we know is that after both candidates were eliminated Frew received 1,539.8 votes, Storey received 648.6 votes and O’Lynn received 2,225.9 votes. There were 974 votes left undistributed. Splitting the stage in two ultimately would not have changed the result but it was odd to do it this way and also means we cannot tell how many of those undistributed votes were TUV votes or indeed how many TUV votes transferred to Alliance instead of the DUP.
There were several stages were multiple candidates were eliminated at the same time so all of these stages have been ignored when showing where a party’s votes went. For the sake of working out percentages, the figures simply have not been included, which I thought made more sense than adding an ‘unknown’ section. It also means there has been no guesswork or assumptions made on these cases, as will be seen below there are often some interesting transfer patterns that seem to make no sense to an observer.
Interesting Cases
I want to highlight some interesting cases of transfers. In East Londonderry Stephanie Quigley stood as an independent candidate. Quigley was eliminated on 1895.8 votes, of those votes 1,115 (61.8) transferred to unionists with Claire Sugden receiving 749 (41.5%). An additional 212 (11.7%) could not be transferred. What makes this stage interesting is that Quigley was standing as an independent nationalist and had previously been a member of the SDLP. The SDLP only received 195 (10.8%) votes with the other nationalist, Kathleen McGurk of Sinn Féin, receiving 104 (5.8) votes. We can only wonder why so many people who gave Quigley a first preference transferred to unionist candidates, but this was certainly one case where I had to triple check the data because it looked so weird.
Another interesting case is that of the Unionist battle in Foyle. The UUP’s Ryan McCready polled 3,744 first preference votes putting him 357 votes behind Gary Middleton of the DUP. After 13 stages of transfers McCready finished on 4,635 votes, just 95.4 votes behind Gary Middleton. One key stage of this count was stage 9 when the Aontú candidate, Emmett Doyle, was eliminated. The DUP received 176 votes to the UUP’s 38.6 in this round. The DUP received 6.5% of all Aontú transfers or 12% of transfers after we exclude the non transferrable votes. This round arguably decided the destination of the final seat so perhaps the DUP have Aontú voters to thank for their 25th Assembly seat.
Strangford is the other example I want to look at. The TUV’s Stephen Cooper received 5,186 (12.7%) first preference votes while Alliance’s second candidate Nick Mathison only received 2,822 (7%) votes. The DUP ran 3 candidates and the UUP ran 2 so this was a packed Unionist field. In every round of transfers Cooper received fewer transfers than the other Unionist candidates. On the final stage Cooper was 249.8 votes behind Mathison while Mike Nesbitt was 533.1 votes ahead of Mathison. Alliance did well to attract so many transfers and the TUV were probably hindered by there being too many Unionists competing for transfers. Being elected on just 7% of the vote, with very few surplus votes from your own party, is a remarkable achievement and certainly worth highlighting especially since there appeared to be enough votes to elect a 4th unionist.
Voter trends
Alliance have traditionally been seen as a soft Unionist party, and while the party is clear that they have no stance on the border issue they draw in many soft unionist voters or unionist voters wanting to vote for a liberal party. Despite this only 15.6% of all Alliance transfers went to Unionists. Part of the reason for this is that Alliance were actively competing with Unionist candidates for several seats, We will also show below that Alliance transferred a lot of votes to the SDLP in constituencies where Alliance probably won over new voters from the SDLP or Sinn Féin.
Alliance did, however, benefit from a significant portion of UUP transfers. Alliance received 16.7% of all UUP transfers and that figure rises to 23.9% when we remove UUP transfers to other UUP candidates. By comparison UUP voters transferred to the DUP at a rate of 22.8% and 32.6% with no UUP and to the TUV at a rate of 16.4% and 19.8% with no UUP. Alliance will be encouraged that, despite winning over many DUP and UUP voters in recent years, they are still winning plenty of transfers from unionist voters.
Before the election I speculated that Alliance might be in with a chance at gaining seats in West Tyrone and possibly Newry and Armagh. The party had done well in local elections and picked up votes from Sinn Féin in the 2019 European election. These gains did not materialise but is there any evidence to suggest more nationalist voters are being drawn towards Alliance?
While Alliance definitely won over some nationalist voters in the east they struggled to do so in the west, but they did do well from nationalist transfers. Alliance received 23.1% of all Sinn Féin transfers (The SDLP received 36.7%) and 19.8% of all SDLP transfers (Sinn Féin received 18.5%). If we remove parties transferring to their running mates then Alliance received 28.6% of Sinn Féin transfers (SDLP received 45.5%) and 24.2% of SDLP transfers (Sinn Féin received 23.9%).
On the flip side 48.2% of all Alliance votes transferred to the SDLP with 9.2% going to Sinn Féin. These numbers rise to 50.2% from the SDLP and 9.6% to Sinn Féin when we remove Alliance transfers to themselves. That is a majority of Alliance votes transferring to nationalists.
There are some caveats here, there were 9 occasions when Alliance votes could transfer to the SDLP. Of the 11,264.4 votes that transferred from Alliance to the SDLP, 8,206.4 were in constituencies that elected a nationalist MP in 2019, and only one of those constituencies has elected a unionist MP since 2001. Of the 2,149.3 votes that transferred from Alliance to Sinn Féin, only 15.8 were transferred in a constituency that has a non-nationalist MP. This stat should be read with caution because Sinn Féin had many of their candidates elected on the first count in places where they only fielded one candidate so there were very few chances for Alliance votes to transfer to Sinn Féin in these areas.
Alliance also have nationalist transfers to thank for at least 3 of their seats with SDLP transfers specifically to thank for another. Alliance were actually outpolled by the UUP in East Antrim but won a second seat comfortably thanks to nationalist transfers. SDLP transfers (some having come from Sinn Féin in previous rounds) also settled the final seats in both Strangford and North Antrim with the DUP and TUV missing out. In Upper Bann 1,722 of Dolores Kelly’s (SDLP) votes transferred to Alliance while only 1,581 transferred to Sinn Féin and those were split between two candidates. In the final round in Upper Bann Alliance received 267 votes from a Sinn Féin surplus while the second Sinn Féin candidate received 464. It is very rare for Sinn Féin voters to do this, I suspect those 267 had previously been transferred to John O’Dowd from the SDLP.
It is clear that nationalist voters are being drawn towards Alliance. Naomi Long polled 18.5% in the 2019 European election and she clearly won nationalist voters then. In 2019 Alliance won seats in Derry City and Strabane Council and, Fermanagh and Omagh Council for the first time, 2 of their gains were from nationalists. Transfer data shows that nationalists are becoming more likely to transfer to Alliance, and not just to stop unionists from winning, many are ranking Alliance ahead of other nationalist candidates. The SDLP may find themselves being squeezed further by Alliance in the future with their seat in East Londonderry being a prime target for Alliance.
Sinn Féin have not suffered directly from the growth of Alliance but their voters are starting to see Alliance as an option. Sinn Féin voters have a reputation for being ‘plumpers’ or not transferring well to other parties. The fact that Alliance received more than 23% of all Sinn Féin transfers is a significant development and this potential movement of voters between those parties is something I will keep an eye on in future elections.
Moving on from Alliance we have to consider the battle between the DUP and TUV. More than half of TUV votes transferred to the DUP. This is huge but essentially confirms that the bulk of the TUV’s vote was made up of disaffected DUP supporters. Interestingly there were also some significant non transfers.
In South down 772.8 TUV votes were non transferrable on the final stage despite one DUP candidate and 2 nationalists remaining. In North Down 488.3 TUV votes did not transfer with 2 unionists and 2 others remaining. In North Belfast 1,207.6 TUV votes were not transferred with one unionist, one nationalist and one other remaining. A further 430 were not transferred on the final stage in East Belfast with 2 unionists and one Green remaining. It is possible that some of these non transferred votes in the final stage were simply not looked at. If the votes were separated into piles and after a certain number had been counted it would have become clear who the winning candidates were. I am not sure if this is the case but it would certainly make sense when we consider the final stage in South Antrim. The official record shows that, Pam Cameron had a surplus of 1,699.2 votes and that her DUP running mate received 145 of those, the SDLP’s Roisin Lynch received 25.1 and a huge 1,529.1 were not transferred. If it was not the case that they only counted until such a stage that Lynch could not catch Clarke then this makes sense, if not then the DUP have some serious work to do to discover why so many of Cameron’s voters did not transfer to Clarke.
Plumpers
Let’s consider the plumpers, voters who vote for one candidate or one party and don’t transfer to anyone else. Nobody is really sure why voters do this, perhaps it is confusion over the electoral system or perhaps it’s a form of protest to show that only one party or candidate is deserving of their vote. It is difficult to measure how much of an effect plumpers actually have on an election. Those votes contribute to working out the quota and higher quotas tend to favour larger parties. It can also lead to a certain bloc not winning the amount of seats their combined votes would suggest they could.
Across all 18 constituencies 2,277.6 first preference voters were non transferrable after the first round. This equates to 6.9% of votes not being transferred after only one stage of counting. Only 5 constituencies eliminated more than one candidate after stage one. This means that in 13 constituencies we can say for certain people only voted for one candidate. So at least 949.6 (3.8%) votes did not express a second preference. The number of voters who did not mark a second preference is higher because many of these voters were fractional votes distributed from a surplus. What is interesting is that quite a few of these were Sinn Féin votes which did not even transfer to the second Sinn Féin runner. I began this section fully expecting to show that Aontú voters were the plumpers but Sinn Féin clearly still have plenty of voters who will only vote for them or indeed only one of their candidates. It does look like a significant number of Aontú voters only voted for the one candidate, but unfortunately I cannot confirm this with the available data.
Is alphabetical order important?
It has often been speculated that appearing higher up the ballot paper is more beneficial for candidates. So is this actually the case? We can look at North Belfast where it took until stage 10 for the first DUP candidate to be elected. In 7 of the 9 rounds of transfers Phillip Brett received more votes than his running mate Brian Kingston. In West Belfast, Pat Sheehan received the fewest transfers of the 3 remaining Sinn Féin candidates (Flynn and Reilly) in 8 of the 11 stages. It is possible that voters back their preferred candidate and then vote down the ballot paper after finding the party they want to transfer to.
I don’t think this is enough of a sample size to make any concrete conclusions but we certainly haven’t shown that alphabetical order is not important. There are of course many other factors that come into to where transfers go within a party. Parties will split areas up and encourage voters to vote for different candidates depending on where they are. This is most likely the reason why Paul Frew overtook Mervyn Storey in North Antrim with voters from the Ballymena being encouraged to rank Frew above Storey, though he does also come before Storey on the ballot paper which is in alphabetical order. Some voters may simply like or dislike certain candidates.
Transfers between the 3 blocs.
Transfers within a designation are to be expected, it should not be a surprise that a majority of unionists transfer to unionists etc. but has anyone been able to make a significant breakthrough with another bloc?
Just 0.5% of all DUP transfers went to nationalists, 3.7% of UUP transfers went to nationalists, only 2.4% of TUV transfers went to nationalists and, just 0.7% of all other Unionist transfers went to nationalists. The non aligned bloc faired better attracting 4.2% of all DUP transfers, 18.5% of all UUP transfers, 3.7% of all TUV transfers and, 14.4% of all other unionist transfers.
Unionists attracted 15% of all alliance transfers, 6.7% of all Green transfers and, 6.5% of all People Before Profit transfers. Nationalists were able to win 57.9% of all Alliance transfers (the SDLP won 49.2%), 28.4% of all Green transfers and, 24.4% of all People Before Profit transfers.
Unionists won just 1.1% of Sinn Féin transfers, 4.9% of all SDLP transfers, 8.6% of all Aontú transfers and 18.1% of all other nationalist transfers (note this is heavily skewed by the above mentioned transfers from Stephanie Quigley in East Londonderry, but also shows a relatively high transfer rate from Aontú to the DUP). The non aligned bloc attracted 29.1% of all SDLP transfers, 28.5% of all Sinn Féin transfers, 14.2% of all Aontú transfers and, 18.7% of all other nationalist transfers.
Unsurprisingly the non aligned bloc proved to be the most capable of attracting transfers from other blocs. Alliance were particularly effective at attracting nationalist transfers late on in the eastern constituencies, particularly Strangford, East Antrim and North Antrim, or rather, the three seats they gained from unionists.
Where every party sent their transfers
I have created a charts to show where every party’s transfers went after exuding transfers to themselves. I can make the full data available on request but here is a much easier way to see a visualisation of the data.
DUP Transfers
TUV Transfers
Other Unionist Transfers
Other Nationalist Transfers
Transfers from a party to another party (from down, to across)
Access to the data
I have looked at a lot of data to write this piece and putting it all in here would be overwhelming. If anyone would like to see the data for themselves I am happy to share it, simply send me a message and I will send you a link to a version I will put online.
Comments
Post a Comment