Analysing the 2023 Local Elections: Part 4: Last Effective Count (LEC); non-transferable votes; who won the last seat and who will be challenging for it in 2027; Bloc votes by council; correlating SF, DUP and Alliance votes; where ‘Middle-NI’ is…

(Part 1 {evolution of the three bloc votes since 1918, bloc/party plurality, unionist decline problem, Alliance plateauing and perhaps decreasing, } can be found here.)

(Part 2 {the myth of unionist vote shredding, turnout, party performance
} can be found here.)

(Part 3 {identity bloc performance, estimated six-county vote, SDLP and UUP decline, spoiled votes
} can be found here.)


Last Effective Count (LEC):

The LEC involves analysing the votes in the last count that were not eliminated. This includes the quota for candidates who were elected and had their surplus distributed, elected candidates whose surplus wasn’t distributed, and candidates battling for the final seat(s).

Why is this worth doing? Low-vote candidates will have been eliminated before the final count. We can see what bloc their vote transferred to.

The last-seat biggest-vote unsuccessful challenger party allegiances were: UUP (15); DUP (14); AP (12); SDLP (12); SF (11); TUV (6); PBP, Aontú, Ind.N. and Ind.U. (2 each); PUP and Greens (1 each).

The official results from two EAs (The Mournes (NM&D) and Balmoral (B)) show an incomplete count. Some of the transfers in the final stage were not counted, giving LEC votes plus NTs less than the total valid poll. I used the second-last stage for these two EAs. Of the 469 stages involving transfers (i.e. all stages except for the 80 stage 1 first-preferences), 467 are used here, which is 99.57% of all counts where transfers were distributed.

There were 549 remaining candidates in the 78 EA LECs and the second-last counts in Balmoral and The Mournes. With an original total of 806 candidates standing, 31.89% of candidates had already been eliminated.

Looking at bloc share first:

Identity Bloc 2023 vote share % 2023 LEC vote share % % difference
Nationalist bloc 44.27 43.94 -0.33
Unionist bloc 40.12 40.53 +0.41
Middle-ground bloc 15.61 15.53 -0.08

The elimination of the approximately one-third of candidates who didn’t make it to the LEC hasn’t significantly changed the identity bloc sizes. This suggests that my original designation of candidates as unionist, nationalist and middle-ground was broadly correct. N44-U40-MG16, then, are the relative sizes of the three main identity blocs.


Looking at party share:

Party or independent candidates 2023 vote share % 2023 LEC vote share % % difference
SF 30.96 31.27 +0.31
DUP 23.15 24.73 +1.58
Alliance 13.31 13.78 +0.47
UUP 10.96 11.59 +0.63
SDLP 8.72 9.33 +0.61
TUV 3.92 2.62 -1.30
Greens 1.70 1.32 -0.38
PBP 1.08 0.78 -0.30
Aontú 0.91 0.27 -0.64
PUP 0.28 0.27 -0.01
Independent N
(incl. WP & IRSP)
2.60 2.29 -0.31
Independent U
(incl. Con.)
1.81 1.32 -0.49
Independent MG
(incl. SP and CCLA)
0.59 0.43 -0.16

The five main parties all increased their LEC vote share. Smaller parties and independent bloc vote shares all declined in the LEC. Small parties, whose vote is concentrated in certain places (e.g. Greens, PBP, PUP), saw a smaller LEC drop than small parties with a more diffuse vote-share geography (e.g. TUV and Aontú). It looks as if a party needs an overall share of at least 8% to be in with a chance of doing better in the LECs. And the LEC matters, because it is the only count that always elects at least one councillor. The LEC bonus carries over into a seats bonus: the four main parties all gained a seats bonus, and the SDLP were one seat short of it.

The TUV had a mixed election compared to 2019. They had nine councillors elected, up three on 2019. Their vote share was 3.92%; their LEC % share was 2.62%; nine seats is 1.95% of the 462 seats. Their LEC vote share drop of -1.3% on their stage 1 vote share is very similar to the DUP’s gain. Not being transfer-friendly does cause problems.


Non-transferable votes:

The number of non-transferable votes was 27,292, which was 3.66% of the total valid poll. As the total valid poll was 745,515, this gives a non-eliminated vote total of 718,223. This non-transferable percentage is comparable to the 3.78% UK-wide gap between Leave and Remain in the Brexit referendum. Pro-UK and pro-UI sides would need to analyse this chunk of voters as a border poll referendum is likely to be close.

It’s difficult to draw any significant conclusions regarding non-transferable (NT) votes. The map below does suggest that non-nationalist EAs tend to have a lower proportion of non-transferable votes, but Lisburn South and the Downshires contradict that. Similarly, Dunsilly and Clogher Valley are nationalist-plurality EAs but have relatively low NT rates. The quasi-random pattern may simply reflect the order of elimination of candidates.


The battle for the last seat: 2023 and 2027:

In over half the EAs, the final seat was filled by a candidate from the same bloc as the highest unsuccessful candidate (successful candidate first: UUP-UUP (6); UUP-DUP (5); SF-SDLP (4); SF-SF (3); DUP-UUP (3); DUP-DUP (2); AP-AP (2) are the largest components here):

Demographic change in favour of the nationalist bloc is likely to continue for the next four years. There are 12 EAs where either a SF or SDLP (six each) candidate was the unsuccessful challenger for the last seat which went to a candidate from either U or MG blocs:

The votes needed for SF/SDLP to win these seats are (sorted by 2023 gap as % of valid poll):

Electoral Area (Council): Last seat victor: Last seat biggest challenger: Vote gap (% of valid poll):
Dunsilly (A&N) Alliance SDLP 21 (0.28%)
Enniskillen (F&O) UUP SF 65 (0.85%)
Causeway (CC&G) DUP SF 85 (0.96%)
Ards Peninsula (A&ND) UUP SF 104 (1.12%)
Ballymena (M&EA) DUP SDLP 85 (1.17%)
Erne North (F&O) UUP SDLP 112 (1.58%)
Killultagh (L&C) UUP SDLP 206 (2.46%)
Lisnasharragh (B) GP SF 331 (2.76%)
Airport (A&N) Alliance SDLP 261 (3.26%)
Botanic (B) GP SF 388 (3.66%)
Carntogher (MU) DUP SDLP 688 (7.79%)
Lagan River (AB&C) DUP SF 853 (8.53%)

Given the large swing towards nationalism in this election, the first ten are likely seat-changes next time. Carntogher and Lagan River would require a vote shift of half a quota to change.


Bloc votes by Council:

These are my estimated bloc votes by council, sorted by valid poll % (highest turnout first):

Council Valid Poll % U bloc % N bloc % MG bloc %
Mid Ulster 64.75 32.05 64.67 3.28
Fermanagh & Omagh 62.09 31.26 61.55 7.19
Newry, Mourne & Down 57.48 18.33 71.28 10.39
Armagh, Banbridge & Craigavon 55.63 50.52 38.94 10.54
Derry City & Strabane 55.61 22.20 73.33 4.47
Causeway Coast & Glens 53.50 50.00 39.36 10.64
Belfast 52.24 28.70 50.53 20.77
Lisburn & Castlereagh 51.31 52.03 17.67 30.30
Antrim & Newtownabbey 50.39 50.72 27.47 21.81
Mid & East Antrim 49.17 68.85 11.31 19.85
Ards & North Down 46.01 59.62 5.37 35.00

To generalise: the higher the turnout, the higher the nationalist bloc share, and the lower the other two bloc shares.


Correlating the three main party performances:

In the three graphs below, the datapoints represent the vote share by two parties in each of the 80 EAs. A linear best-fit line (using Excel) was drawn, and the R-squared value for that best-fit line is displayed. (The higher the R-squared value, the greater the correlation. The maximum possible value is 1, when all points would lie on the line.)

The highest correlation is between SF and Alliance performance: where SF is strong, Alliance is weak; and vice-versa. Translating this into border poll referendum campaigning: pro-unity arguments that resonate with SF voters (one-third of the electorate) may well have the opposite effect in the Alliance heartland. Conversely, constructing pro-unity arguments that appeal to Alliance voters (e.g. keeping a devolved Assembly) may not resonate with border nationalists.


Where is ‘Middle-NI’?

Warning: statistics discussion!

What part of Northern Ireland is most typical, in terms of voting pattern? We can answer this question by calculating how far the N, U and MG bloc percentages deviate from the overall bloc shares. It’s not enough to add the three differences because the sum of these would give 0. For example, using the N44-U40-MG16 overall bloc shares, an EA with 42-41-17 has differences of -2+1+1=0. However, if we square the differences, we get 4+1+1=6. If another EA has 47-42-11, the squares of the differences are 9+4+25=38. So the bigger the number, the more the EA deviates from the overall vote share. After the squares are added together, the square root of this sum was calculated: this is called the standard deviation, and is the figure used in the two maps below.

Statistics discussion finished!

Sorry about the statistics! To answer the question: Enniskillen most closely approximated the overall bloc vote share, just ahead of Craigavon. Arguably, Craigavon’s is more typical as the N bloc share is just ahead of the U bloc, whereas the opposite is the case in Enniskillen.


Regarding the EA whose party result was most typical of NI, Dunsilly EA wins that accolade, closely followed by neighbouring Airport and Glengormley Urban.


South Antrim residents can expect an influx of reporters come election or referendum time. It isn’t easy being archetypical!


Discover more from Slugger O'Toole

Subscribe to get the latest posts to your email.

We are reader supported. Donate to keep Slugger lit!

For over 20 years, Slugger has been an independent place for debate and new ideas. We have published over 40,000 posts and over one and a half million comments on the site. Each month we have over 70,000 readers. All this we have accomplished with only volunteers we have never had any paid staff.

Slugger does not receive any funding, and we respect our readers, so we will never run intrusive ads or sponsored posts. Instead, we are reader-supported. Help us keep Slugger independent by becoming a friend of Slugger. While we run a tight ship and no one gets paid to write, we need money to help us cover our costs.

If you like what we do, we are asking you to consider giving a monthly donation of any amount, or you can give a one-off donation. Any amount is appreciated.