The Turning Point? The Northern Ireland Local Government Elections of 18 May 2023

The Northern Ireland Local Government elections of 18 May 2023, merit the slightly overused accolade of “historic.” More cautious observers may see them as a potential “inflexion point.”

Brendan O’Leary and Jamie Pow

The Northern Ireland Local Government elections of 18 May 2023, merit the slightly overused accolade of “historic.” More cautious observers may see them as a potential “inflexion point.”

There are eleven local government districts in Northern Ireland from which 462 councillors are elected. Eight districts choose 40 councillors; two elect 41 each (Newry, Mourne and Down, and Armagh, Banbridge & Craigavon), while Belfast chooses 60. The electoral system is the single transferable vote (STV) system of proportional representation. The ‘district electoral area’ (DEA) is the unit from which multiple councillors are elected. Notoriously Northern Ireland local governments have few powers so elections to them are widely seen as exercises in displaying the strength of the region’s political parties and blocs.

These elections had been postponed for two weeks to avoid a clash with the coronation of Charles III. There was a region-wide turnout of 54%, a rise of two percentage points on the previous local elections in May 2019. By comparison, turnout for local government elections in the Republic was 50% in 2019. The results in the North emphatically underline recent electoral shifts, underpinned by demographic change, and the outworkings of “Brexit.” They are also consistent with the 2022 May Northern Ireland Assembly elections previously reported here.

Confirming demographic and electoral change

For the first time in Northern local government elections, the first-preference votes (FPV) of those who backed pro-unification parties or candidates unambiguously outnumbered those who backed pro-union parties and candidates: by 43.8 to 39.9 per cent. See Table 1. And for the first time in this form of election the entire pro-union bloc fell just below 40 per cent. Differently put, in this election there was a 44:40:16 breakdown on the national question among nationalists, unionists, and others respectively. The 40:40:20 description habitual in recent commentary may need to be updated.

Table 1. First-preference vote in  2023 local government elections  by bloc

Sources: Data from Local Government Districts websites. Calculations regarding independents explained in the text by the authors.

Was there differential turnout by bloc? In our 2022 Assembly election survey, turnout rates among those identifying as nationalist and unionist respectively were practically identical (at NI level); it was among the ‘neithers’ that turnout was significantly lower. Likewise, Protestant and Catholic turnout in 2022 was virtually identical. So, did differential turnout matter this time?

Local  government elections always have a lower turnout than Assembly or Westminster elections, but, in this case there was an increased turnout of two percentage points on 2019. In Table 2 below we compare raw FPV totals in the 2022 Assembly and the 2023 Local Government elections. The comparison comes with two qualifications: it is  not always wise to make direct comparisons between different forms of elections; and  the eligible electorate changed slightly  within the year. But given the relative proximity of the two contests, it is noteworthy that the 2023 FPV totals for the main nationalist parties represent just over 90% of the number of votes they received in May 2022, whereas  unionist parties had a significantly lower percentage,  82%, suggesting an aggregate mobilization and enthusiasm gap between the two blocs, though we accept the picture at district level is more complex.

Table 2. First-preference vote by bloc in 2022 Assembly and 2023 Local Government elections

Sources: Data from Local Government Districts websites. Calculations regarding independents explained in the text by the authors.

Calculating bloc totals. Expected designation of party or independents as nationalist, unionist, or other is not formally pertinent in Northern Ireland local government elections as it is in the Assembly. But parties and independents campaign strongly as players within these three blocs. The authors independently examined all the parties’ and independent candidates’ formal stance on the national question to address whether they favour Irish unification, maintaining the Union, or are neutral, indifferent, prefer to postpone, or are simply uninterested in “the national question”—in which case they are coded as “other.”  The results of our codings are in Table 1.

Parties. The codings of parties should be uncontroversial. The “unionist parties”, or big-U parties,  are the Democratic Unionist Party (DUP), the Ulster Unionist Party (UUP), Traditional Unionist Voice (TUV), the Progressive Unionist Party (PUP), and the Northern Ireland Conservatives.  The “nationalist parties” are Sinn Féin, the Social Democratic and Labour Party (SDLP), Aontú, People before Profit (PBP), and the Irish Republican Socialist Party (IRSP). The PBP would designate as “other” in the Assembly, but because it favours a united Ireland, we code it as part of the nationalist bloc. The “other parties” are the Alliance, the Greens, Labour Alternative, the Workers’ Party, and the Socialist Party. The last three named micro-socialist parties are fully “other” in the sense that they either avoid the national question or take no formal position on Irish unity or the union—yet. We mildly disagree with Dr Philip McGuinness’s coding of the Workers’ Party in his recent illuminating articles on Slugger O’Toole: he includes that party in the nationalist bloc. The Workers’ party’s website, albeit after the elections, condemns both nationalisms (British and Irish) and seeks to focus workers’ unity on class questions, bracketing the national question—for now. So, we code it as “other.”

Independents. The coding of independents is where the trickiest questions may be expected to arise; after all, they call themselves independents. We independently coded each independent candidate as nationalist or unionist or other, judging each person by their election statements, if available, or by information on the web. We independently agreed on the coding of all the independents without any need for discussion of “hard cases.” To take one example to illustrate our procedures, we coded Tony Mallon, who stood in West Belfast,  as “other” because he favours an independent Northern Ireland. Sceptics about our codings are invited to check the attached file, which embeds one web reference for each independent: many further web-based references could have been provided. See document entitled ‘NI election 2023 votes’ in the Downloads section of this page.

Leading party performances in each bloc. Sinn Féin emphatically won first place across the region—both in first-preference votes (31%, a gain of 7.8% on the previous local government election when it had won 23.2%), and in overall council seats won (144, a gain of 39, 31.2%). Sinn Féin thereby decisively displaced the DUP from its previous perch as the largest party in local government. Sinn Féin’s percentage of the FPV was an improvement on its surge in the Assembly elections of 2022. Its gains were partly at the expense of the SDLP, but judging by recent surveys Sinn Féin seems to be gaining support through dominance among the newest and youngest voters. A pre-election poll by the Irish News and the Institute of Irish Studies at Liverpool University showed Sinn Féin had 36.5% of the intended first preference vote among those aged under 25. In the ARINS-Irish Times survey, partly reported in the Irish Times in December 2022 and January 2023 and to be archived on this website, the directly comparable figure is 42.4 percent.*

The DUP came second in the region in first-preference votes (23.2%), a slight reduction in its vote-share in the last local government elections (24.1%). But it eked out exactly the same number of councillors (122), with net losses and gains exactly cancelling one another out. While Sinn Féin’s FPV vote and seat share were almost identical, the DUP had a seat bonus (winning 26.4% of the total councillors) largely because it benefitted from transfers from other unionist parties. The DUP set its target low, keeping its seat share, and aimed to stop inroads into its base by the hardline TUV, which held the DUP responsible for “the sea border” created by the Protocol. The TUV increased its number of councillors (from 6 to 9), but won just 3.9% of the FPV, by comparison with 7.6% it had won in the Assembly elections a year earlier (although this may in part reflect the lack of TUV candidates standing in the majority of DEAs in this year’s local elections).

The Alliance Party came third in first-preference votes (13.3%), falling just shy of 100,000 votes, and it won 67 council seats (67/462 = 14.5%), a small seat bonus through lower-order transfers from voters for other parties, but a net gain of 14 councillors since 2019. (The party is appealing the electoral count in Derry and Strabane, where a counting error may have cost it a seat in favour of the SDLP.) Alliance remains weak in western and southern Northern Ireland, indeed in all electoral districts adjacent to the border with the Republic. Alliance’s modest success was partly at the expense of the Green Party. Alliance is formally neutral on the question of whether Northern Ireland should remain in the United Kingdom or reunify with the rest of Ireland.

Analysis of electoral data over time by Dr Philip McGuinness suggests that “Alliance’s best days may be behind them.” That remains to be seen, but going forward, rather than seeking equidistance, Alliance may have to lean further toward the nationalist bloc than the unionist bloc. On the day before the local government elections, Dr Sydney Elliott, analysing patterns of transfers from defeated Alliance candidates in recent elections, observed in the unionist paper, the NewsLetter that

Alliance terminal transfers [in the three STV-PR elections held between 2017 and 2022] have gone increasingly to SDLP, SF and other republican parties. The difference in preference over unionist parties rose from 5.1% in 2017 to 18.2% in 2019 and 48.3% in 2022. At face value, this indicates that modern Alliance voters are more inclined towards SDLP, SF and other nationalist and republican parties. This trend may be influenced by Alliance seeking to extend outwards from its greater Belfast base to traditionally more nationalist areas where there may be fewer unionist party candidates.

The second-placed in the two major blocs. The soft nationalists of the SDLP (39 seats, down -20) and the soft unionists of the UUP (54 seats, down -21) suffered almost identical losses in numbers of seats, though the percentage share lost by the SDLP was higher (-34%) compared to the UPP (-28%). The SDLP and the UUP were the parties which negotiated the Good Friday Agreement (GFA) as the leading parties in their blocs. They are now down to 8.4% and 11.7% respectively of the total council seats available, with no seat bonus for the SDLP (it won 8.7% of FPVs) and a marginal one for the UUP (with 10.9% of FPVs). The memberships of these two parties are regularly described as ageing, like their most reliable voters, even though their leaderships have been renewed. It is difficult to see how either party can reverse its decline.

Past the tipping point. Northern Ireland was invented in the UK’s Government of Ireland Act of 1920, but whatever the exact definition of its formation it  has just had its centennial. Designed to encompass the entirety of what Edward Carson had called “the six plantation counties,” its borders were selected to ensure a two-to-one Protestant and Unionist majority. The 2023 local government elections display neither a Protestant demographic majority nor a Unionist electoral majority: a century after its formation the founding rationale of Northern Ireland has faded into the twilight. According to Dr Philip McGuinness, in the May 2023 local government elections unionist parties and independent candidates with a capital U jointly won a majority of the first preference vote in none of the six counties. In four counties nationalists won a majority of the vote: Fermanagh, Tyrone, Armagh, and Londonderry. Dr McGuinness’s minor differences with us over how to count the nationalist bloc would make little difference to this assessment.

In just two counties are unionists the plurality bloc: Antrim and Down. In the two leading cities Derry returned its customary nationalist majority while in Belfast the nationalist bloc, by our coding, took 49.6% of the first preference vote, which would be 50% on standard rounding. McGuinness has Belfast as a nationalist majority city but that is because he codes the Workers’ Party as nationalist. McGuinness has also shown that nationalists now outnumber unionists as the plurality bloc in 40 DEAs, compared with 34 where unionists hold that status, and 6 where the others are the leaders. The respective standings of the two leading blocs have therefore almost exactly reversed since the last elections when unionists were the plurality bloc in 40 DEAs and nationalists in 35.

The outcomes across councils

There are now four safe nationalist councils, in contrast with two safe unionist councils—in which unionists with a capital U have a combined majority. See Table 3.

Table 3. 2023 Local Government Elections in Northern Ireland. Councillors elected by the two major blocs (%) are displayed  in order of strength of nationalist representation.

NOTES: In this table, the nationalist bloc on our codings = Sinn Féin, SDLP, Aontú, SPBP, while the unionist bloc = DUP, UUP, TUV, PUP.

The number in parenthesis references the relevant council district displayed in Figure 1 below.

In Figure 1 below, created by Philip McGuinness, the gap between the nationalists and unionist blocs in the first preference-vote is shown for each local government district. Five councils lack either a nationalist or a unionist, with a capital U, majority. In Belfast, nationalists are the leading bloc, close to obtaining a majority, and look likely to obtain such a majority five years from now. In Armagh, Banbridge and Craigavon, unionists hold that position, but seem likely to slip a little in future. In three councils unionists with a capital U comprise fifty percent of the councillors (Causeway Coast & Glens, Antrim & Newtownabbey, and Lisburn & Castlereagh). They have plurality status, which they may retain in future elections even if they slip below 50%.

© Dr Philip McGuinness, map reproduced with kind permission.

Figure 1. The difference between first preference votes for the nationalist bloc and the unionist bloc in each of the 11 Local Government Districts.

Note: The numbers in the first column of Table 3 refer to the council districts displayed in this map.

The underlying political sociology. Two underlying factors likely explain the tipping or inflection point marked by these elections: demographic and Brexit effects.

Demographic effects. According to the Northern Ireland census of 2021, partly published in 2022, cultural Catholics—those who identify as Catholics or who were brought up in Catholic households—outnumbered cultural Protestants for the first time since a census was counted in the six counties that became Northern Ireland—see Figure 2.

Figure 2. Northern Ireland’s religious demography: Proportions, 1861-2021

NOTES

The graphic is © The Authors, reproducible with acknowledgment. Sources: Census of Ireland 1861-1911, Northern Ireland Census 1926-2021

(i) The 2001, 2011, and 2021  census reports assign many of the ‘none’ and ‘not stated’ to ‘community background’ or ‘religion brought up in’ (based on answers given by respondents) and they are then classified as ‘Catholic’ or ‘Protestant’ as appropriate, along with those who identify as such. So, from 2001 the Protestant and Catholic lines exaggerate formal religious identification & and cannot be taken as a guide to the proportions of believing or practicing Catholics or Protestants. (ii) The 1981 census was subject to a partial boycott, by Catholics. (iii) Red line: Not stated, was discontinued in 2001. (iv) Yellow line: Nones, were 9.3% in 2021.

No one denies the strongly established link between demography and electoral behavior in Northern Ireland: most unionists have been cultural Protestants, and most nationalists have been cultural Catholics. The scope of the effects is disputed, and the degree to which they are changing. For now, recent ARINS-Irish Times survey data confirm that the proportion of Catholics who favour the maintenance of the Union is higher than the proportion of Protestants who favour Irish reunification.

The absolute numbers of those identifying explicitly with a religion in 2021 are shown in Figure 3. Among religious adherents, a Catholic demographic majority has arrived. Other non-census indicators, not reported here, suggest that all major Christian denominations are experiencing fast-paced secularization.

Figure 3. Northern Ireland’s religious demography: Absolute numbers of adherents 1861-2021

NOTES

The graphic is © The Authors, reproducible with acknowledgment.

(i) There are now absolutely more Catholics than all Protestants and other Christians and the religious (805, 151 > 736, 515), and the Catholic population looks set to grow further. (ii) “One million Protestants” in the North is not an accurate shorthand; three quarters of a million is. (iii) None/not stated = fastest rising category since 1991 (iv) During  1891-1937 Presbyterians were  poised to exceed Catholics in the six counties, but their numbers peaked in 1961. Presbyterians  are now 39.3% of Catholic numbers, 77,000 less than at the formation of Northern Ireland, and on a downward trajectory. (v) Church of Ireland numbers peaked in 1951, and are now 27.3% of Catholic numbers, and on a downward trajectory. (vi) Other religions here include other Protestants and Christian religions (e.g., evangelical Protestants, and Quakers). (vii) Over 203,000 Presbyterians and Church of Ireland net losses (1971-2021) are mostly explained by the rise of other Christians (c. 194,000 in this period). (viii) The Free Presbyterian Church founded by Ian Paisley still had less than 10,000 adherents in 2021.

Less clear is whether a cultural Catholic electoral majority will exist before the next census in 2031. To predict that accurately we would need to know the exact breakdown of cultural Catholics among the “Nones.” In Figure 4 we chart the cumulative cultural Catholic demographic lead over cultural Protestants and other Christians according to the broad age cohorts recently published by the census authorities. The curve rises strongly among the young before fading among the over-65s.

Figure 4. Cumulative cultural Catholic demographic lead over cultural Protestants & other Christians in 2021

 

Source of data: Northern Ireland Census 2021

By the end of 2031 all those aged between 8 and 17 in 2021 will have joined the electorate (unless they have died prematurely or emigrated), provided they are eligible to vote. The maturation of this cohort during the rest of this decade will be decisive: in 2021, in the entire age-range from 15 upward, cultural Protestants retained a net demographic advantage over cultural Catholics of roughly 17,000 people. But over the 2020s, that lead will be fully eroded by the steady arrival in the electorate of the cultural Catholics aged 8-17 in 2021. That will complete the demographic reversal in the electorate. What can be said with strong confidence, barring major off-trend migratory movements or a major shift in differential outbound migration at the expense of cultural Catholics, is that by 2030 the electoral fate of Northern Ireland’s future will no longer be in Protestant hands.

These local government election results are therefore being  used to bolster the claims of those who expect a future nationalist majority and a future majority for reunification in referendums that could be initiated under the provisions of the Good Friday Agreement. That may well be the  direction of travel, but current calls for a referendum are certainly premature.

Brexit effects. Another recent transformative driver of regional politics has been “Brexit/UKexit.” Since the UK-wide  referendum of  2016 the traditional national question has been overlaid by the European question. Northern Ireland voted to remain by a margin of 56: 44 % in 2016. In the referendum campaign all the major Northern Irish parties, except the DUP, had favoured “remain.” After the referendum the UUP changed its position, accepting the leave result across the UK. Though some Catholics voted “leave” in 2016, especially in older age categories, Protestants did so far more decisively. A minority of more liberal Protestants, usually voters for Alliance, the Greens, or the UUP, voted to remain.

Since 2017 the combined (pro-remain) nationalist and (pro-remain) “other” vote has usually been stable at or around 55% in region-wide elections in Northern Ireland—the last European Parliamentary elections, Westminster elections, Northern Assembly elections, and local government elections. If we treat the nationalist, unionist, and other blocs as internally homogeneous on the European question then the 2023 local government elections would suggest that the remain bloc now sits at 60.1 percent. Excluding “lexiteers” among some independent nationalists and micro-socialist candidates in the PBP and some of the independent “others”  would reduce that figure. On the other hand, counterfactually, if there were to be another UK referendum on returning to the EU, we would expect the UUP voters to split.

The European question aligned Alliance and the Greens with the nationalist parties, and indirectly this question, in the form of the Protocol, with a new coat of paint in the form of “The Windsor Framework,”  dominated the 2023 local government elections. That is because since its initial drafting, the Ireland/Northern Ireland Protocol—the legal instrument through which Northern Ireland remains in the European Single Market for goods and agriculture, and through which the UK administers the European Union’s customs code at ports and airports between Great Britain and Northern Ireland—has polarized the political parties along traditional nationalist versus unionist lines, with the important exception that the pro-European Alliance and Green parties support the Protocol, and the Windsor Framework.

Since  early 2022 the DUP has boycotted the Northern Ireland Assembly, in protest at the Protocol, thereby disabling two strands of the Good Friday Agreement—the power-sharing institutions within the North, and the North-South Ministerial Council (NSMC). The party took this stance to demand that the Protocol be renegotiated—so that Northern Ireland’s place within the Union and the UK’s internal market would be secure—and because it feared being outflanked by Jim Allister’s TUV—Allister, a former MEP, had been a leading light in the DUP. TUV supporters held the DUP culpable for the Protocol, and for being “Lundies”—traitors to the unionist cause.

Any renegotiation of the Protocol, however, has to occur though its signatories: the UK government and the European Union. The Conservative and Unionist Party, in power at Westminster since 2010, has gone through five Prime Ministers since 2016, and almost as numerous strategic shifts over how to leave the European Union. But since Rishi Sunak became Prime Minister in October 2022 the Conservatives have prioritized improved relations with the European Union. The Windsor Framework was the result, achieved in late February 2023. Unionist lawyers, and possibly theologians, will dispute the extent to which the Framework meets the seven tests that the DUP set for the UK government. Some of them, however, could not be met without terminating the Protocol. Not accepting the Windsor Framework, and demanding further changes to the Protocol, may have helped the DUP stave off the TUV in these local government elections, but it is currently unclear what face-saving formula will enable the party to participate in the Assembly and its power-sharing executive—if it wants to do so.

Four scenarios reduced to two. In early 2022 it was possible to imagine four future institutional scenarios in and over Northern Ireland covering the rest of 2020s. See Figure 5. Two dimensions build these four scenarios. In the first, the question would be whether the Northern Ireland Assembly, Executive, and the North South Ministerial Council (NSMC) function (with or without minor emendations), or not. In the second, the question would be whether the Protocol, as modified by EU-UK negotiations, as in the Windsor Framework, functions or not.

Figure 5. Scenarios for the Good Friday Agreement Institutions and the Protocol 2022-2030

Source: adapted from O’Leary, B. Making Sense of a United Ireland (2022)

In Scenario 1, Glide Path, both sets of institutional arrangements function for the rest of this decade, with referendums on unification called in 2031, or shortly after.

In Scenario 2, Rocky Road, by contrast, the Protocol functions, but the Northern institutions and the NSMC do not—because the DUP refuses to re-enter into the Assembly and Executive and demands further modifications to the Windsor Framework, which neither the London government nor the European Union are willing to accept. In these circumstances a straightforward return to simple direct rule would be rejected by the Government of Ireland, which, like the rest of the EU, will observe that the UK is obliged by its recent treaties with the EU to protect the GFA “in all its parts” and “in all its dimensions.” In this scenario, UK direct rule would  be modified by input from the Government of Ireland through the British-Irish Intergovernmental Conference (BIIGC) which “subsumes” the conference established under the Anglo-Irish Agreement of 1985. The former leader of Alliance, and the first Speaker of the Northern Ireland Assembly, John Alderdice, has labelled this scenario “de facto joint authority.”

We believe that scenarios 3 and 4, logical possibilities in 2021-2, are now exceedingly unlikely. That is because both the UK government (and its likeliest successor, a Labour, or Labour-led government), and the EU are determined to make the Protocol as modified by the Windsor Framework function. That reduces the four scenarios to two, with the DUP in effect deciding which of these will emerge, though it is possible to imagine oscillation between scenarios 1 and 2. We believe that the EU-UK détente, supported by the US, has cut off scenarios 3 and 4 through the Windsor Framework, and that the local government elections of May 2023 have cut off the loyalist hope that they could build sufficient electoral support to overthrow the Protocol.

Notes:

* The authors thank John Garry, Joanne McEvoy, and Philip McGuinness for their assistance and critical commentary.

** Thanks to John Garry. This figure was generated while weighting for ‘past’ vote, i.e., the vote that had taken  place in the May 2022 Assembly elections. The question posed was: if there was a further Assembly election who would you vote for?  The full figures are: 18-24 = 42.4%; 25-34 = 28%; 35-49=27.3%; 50-64=37.2% and 65+=27.1 %.

The Authors:

Brendan O’Leary MRIA (hon), is Lauder Professor of Political Science, University of Pennsylvania, and Honorary Professor of Political Science Queen’s University Belfast. He was recently conferred with an Honorary Doctorate of Arts by University College Cork particularly for his work on Northern Ireland. He is the author of A Treatise on Northern Ireland (3 volumes) (2020), and Making Sense of a United Ireland (2022).

James Pow is Lecturer in the School of History, Anthropology, Philosophy and Politics, Queen’s University Belfast. His research addresses how citizens interact with democratic decision-making, e.g., through elections, mini-publics, and referendums, typically with a focus on political attitudes and behaviour in Northern Ireland. He is an affiliate of the Democratic Innovations and Legitimacy Research Group at KU Leuven, Belgium.

The authors will be among the lead investigators on the next ARINS-Irish Times surveys.