Working towards Irish Unity




The 2022 Northern Ireland Assembly Elections 

Originally published on LSE blog.

The Northern Ireland Assembly election of May 2022 was transformational, despite small shifts in the magnitudes of seats won by the two dominant blocs. John GarryBrendan O’Leary and James Pow discuss the results. 

On 5 May Sinn Féin won first place in the NI Assembly elections – both in first-preference votes (29%), and in seats won (27 of the 90 MLA positions in play, i.e., 30%). As reported across the world, that is the first time any Irish nationalist party has headed the polls in region-wide elections in Northern Ireland, earning Sinn Féin the right to nominate the First Minister. In 1922, the Northern Ireland Parliament had voted to secede from the Irish Free State under Article 12 of the Anglo-Irish Treaty. Northern Ireland’s 100th birthday has therefore coincided with the reversal of its founding rationale. A novel political entity, forged to secure an Ulster unionist and Protestant demographic and electoral majority, no longer has either.

The Democratic Unionist Party (the DUP) came second in first-preference votes (21.3%), a significant loss compared with what it had won in 2017 (28.1%). But the DUP lost just three seats compared to 2017, winning 25 seats (25/90 = 27.8%). This disproportional outcome occurred because its candidates benefitted from lower-order vote-transfers from those who had backed the über-hardline Traditional Unionist Voice (TUV), which had campaigned with “No Sea Border” beside its name on the ballot paper, and won 7.6% of the first-preference vote. The TUV won just one seat (1/90=1.1%), returning its party leader, because its candidates received very few transfers from those who had voted for other unionist parties’ candidates.

The Alliance Party came third in first-preference votes (13.5%), but won 17 seats (17/90 = 18.8%). The party more than doubled its seat-share compared to 2017. It disproportionally benefitted from lower-order transfers from other parties. Its success was partly at the expense of the Green Party, which lost the two MLAs it had won in 2017. Alliance is formally neutral on the question of whether Northern Ireland should remain in the United Kingdom or reunify with the rest of Ireland.

The headline-grabbing achievements of Sinn Féin and Alliance should not obscure the DUP’s robust performance in seats-won. Despite the battering the party took to its standing and prestige before the election it stopped the bleeding – at least for now.


Overall, the election results are suggestive of stability, and of incremental rather than major change, despite the transformational symbolism of Sinn Féin’s win. Stability is evident in four features of the election:

  • Sinn Féin won the same number of seats as it did in 2017 on a marginally improved first-preference vote total.
  • There was evidence that both Sinn Féin and the DUP continued to appeal to be, and to benefit from being seen as, the champions or tribunes of their respective blocs.
  • Northern Ireland continues to have two similarly sized nationalist and unionist blocs: 35-37 in seats, and 41-42% of the first-preference votes—after adding in votes for known nationalist and unionist micro-parties and candidates.
  • The others, the ‘non-aligned’ bloc, remain firmly behind the two major blocs, having 16.4 % of the first-preference votes, but they won 20% of the seats because of Alliance’s ‘seat-bonus’.

Incremental change, however, is reflected in three distinct features of the election:

  • Sinn Féin grew further at the expense of the SDLP within the nationalist bloc, which overall grew slightly compared to its first-preference vote total in 2017 (adding in the micro-nationalist and republican parties and candidates).
  • The DUP faced much greater competition on its right flank (from the TUV) and from its comparatively liberal rival, the Ulster Unionist Party (UUP).
  • The ‘other bloc’ in the Assembly is now dominated by one party – Alliance. It has become the tribune of the others, and the champion of changing Northern Ireland’s consociational arrangements. Alliance’s success is also owed, at the margin, to its success in winning transfers from the soft nationalists and soft unionists, the Social Democratic and Labour Party (SDLP) and the UUP. Alliance is over-represented in the Assembly, and the other bloc also has more seats than its first-preference vote-share. Such data will form part of predictable replies to its complaint that the consociational system is stacked against cross-community parties.

Cross-bloc comparisons

There are different ways to classify the outcome by blocs. By seats won is one method. Table 1 below displays the expected designation-choice of parties or MLAs in the Northern Ireland Assembly. There are 37 unionist MLAs (25 DUP + 9 UUP + 1 TUV + Alex Easton in North Down and Claire Sugden in Londonderry East), 35 nationalists (27 Sinn Féin + 8 SDLP MLAs), and 18 others (17 Alliance + People Before Profit’s Gerry Carroll in West Belfast). Expressed in percentages, the Assembly is 41.1% unionist, 38.9% nationalist, and 20% other.

A second method of bloc-comparison is to present all first-preference votes. Instead of taking expected designation in the Assembly as decisive in classifying a party or a candidate, observers may examine parties’ and independent candidates’ formal stance on the national question. In each, the question to be answered is whether parties or candidates favour Irish reunification, maintaining the Union, or are they neutral, indifferent, prefer to postpone, or are simply uninterested in “the national question” – in which case they are ‘other.’ The results of our codings are in Table 2.

A few words on party codings (see Table 3). Coding the UUP, the DUP, the TUV, and the PUP as unionist was easy: unionist is in their titles. We then added the Conservative Party of Northern Ireland and the Heritage Party to the unionist bloc. Classifying the SDLP, Sinn Féin, Aontú, and the Irish Republican Socialist Party (IRSP) as nationalist was equally easy. These parties favour Irish reunification.

But we also coded the 9, 798 first-preference votes for People Before Profit as nationalist. It is an all-island party, explicitly favours Irish reunification, and explicitly calls for “a border poll” (its words, not ours) in its manifesto. We made this coding decision knowing that in the past the party’s MLAs have chosen to designate in the Assembly as other. Our judgment is reinforced because, despite its formal cross-community appeal, the PBP’s electoral strength, such as it is, is confined to republican heartlands: its prime competition is with Sinn Féin.

The other bloc was also relatively straightforward to code. Alliance and the Greens are the largest parties in this bloc: the former is neutral on the Union; the latter does not prioritise reunification, though it is an all-island party. The same applies to the micro-socialist entities: the Workers’ Party, the Socialist Party, and the Cross Community Labour Alternative (CCLA).

Now some brief words on the coding of independents, where the trickiest questions were expected to arise; after all, they call themselves independents. O’Leary and Pow independently coded each independent candidate as nationalist or unionist or other, judging each of them by their election statements, if available, or by information on the web. Two unionist independents are uncontroversial: Easton (North Down) and Sugden (Londonderry east). We easily agreed that Hynds in Lagan Valley and Moran in South Antrim were independent unionists. We classified four independents as nationalists: Quigley, ex-SDLP, in Londonderry East, a former SDLP councillor who left the party over abortion, McCloskey in Foyle, who has firm republican convictions but is anti-abortion, Gallagher in west Tyrone who is palpably a republican, and DeSouza in Fermanagh and West Tyrone (a slightly more difficult decision). All the remaining independents were classified as other without equivocation on the part of either coder.

The gap between the top two blocs by party-candidates was 3,842 in favour of nationalists. In total, however, after coding and adding in the first-preference votes for independents, the unionist bloc had a net advantage of 6,416 votes (less than a percentage point). Unionist hegemony is over, the unionist plurality survives by a thread.

A power-sharing government: now, later, or not at all?

According to the rules of the Good Friday Agreement, as modified at St Andrews, the party that gets the most seats is entitled to nominate the First Minister in the power-sharing executive. Although the First and deputy First Minister are legally equal, if the DUP refuses to nominate a deputy First Minister, executive formation cannot be completed.

If executive formation does proceed, then most of the rest of the executive will be filled by the d’Hondt rule: see table 4 below. The SDLP will not be entitled to any portfolios.

According to current legislation, Northern Ireland has seven departments headed by a minister, to be filled by the d’Hondt rule. If all parties take the portfolios to which they are entitled, then, as the last row of table 4 shows, there will be three nationalist ministers, three unionist ministers (two DUP, one UUP), and one Alliance minister, allocated through d’Hondt. But, if we add the two First Ministers, and the Justice Minister, who is elected by cross-community consent and almost invariably an Alliance nominee, the final executive composition would be as follows: Sinn Féin 4, DUP 3, UUP 1, and Alliance 2 ministries. Differently put, there would be four nationalists, four unionists, and two others in the cabinet. The Alliance will have 2/10 portfolios, a proportion above both its first-preference vote share and its seat-share, so it can hardly complain that the system works against the party in this respect.

If the parties succeed in their eventual negotiations on government-formation, they might, of course, choose to agree to modify the number of ministers in the executive, but that would require legislation at Westminster, as the current number of ministerial departments is fixed by law, and we’re a long way from that possibility.

Matters of Protocol

The DUP’s message to the UK government, as well as to the electorate, has been that unless the Protocol is scrapped, or significantly amended, the DUP will prohibit the creation of an executive. It says: it’s either the Protocol or the executive but not both.

Although the Northern Ireland Assembly has no legislative power over the Protocol, the MLAs just elected would have a vote at the end of 2024 on whether to discontinue key elements of the Protocol. The Unionist parties are strongly opposed to the Protocol, but their combined seat share is significantly short of the Assembly majority needed to discontinue the relevant Protocol articles: just 37 MLAs out of 90.

The Johnson government simply cannot interpret the Assembly elections as indicating that a majority in Northern Ireland is against the Protocol. Exactly the contrary is the case. Judging by both first-preference votes and seats the majority that favours a functional Protocol has increased compared to the previous Assembly. And unionists are not of one mind on the Protocol: many in the UUP favour a more moderate approach, reform rather than abolition.

Possible referendum on Irish reunification?

Examining the combined seat share of nationalist parties is also important. It provides a very rough proxy of the extent of support for a united Ireland – though some Alliance voters, many Greens, and most People Before Profit supporters would vote for Irish unification in the event of a referendum on the subject. While Sinn Féin retained its 27 seats, the SDLP had a poor election, declining from 12 to 8 seats, losing votes to both Sinn Féin and Aontú, a micro-nationalist party that is against abortion (the SDLP regards it as a matter of conscience for each MLA). The combined nationalist first-preference vote, at 41.4 % by our coding, is marginally behind the unionist vote-share. The Secretary of State for Northern Ireland is unlikely to be obliged to interpret the Assembly elections as indicating a groundswell of support for Irish unity. We are therefore not likely to see an imminent holding of a referendum on the question. That will be deferred until at least the next Assembly.

Another election?

It is quite possible that this May’s election will be looked back upon as the first Assembly election of 2022. Unless the UK government and the EU make an agreement on the running of the Protocol andthe DUP declares itself content with that agreement, we may be headed toward prolonged negotiations to establish a functioning executive. The parties have six months to try to do so, at which point the UK government, according to current legislation, would call a fresh election. So as the parties open discussions to try to reach an agreement on executive formation, their negotiations will be conducted with half an eye to how party positions are likely to play with the electorate in six months’ time.

The DUP may feel obliged to maintain a hard-line position because they perceive their main electoral threat to be the staunchly anti-Protocol TUV, which enjoyed an increase in vote share in this election (though it only retained its one seat). If so, the electoral count centres may have to gear up for a late autumn return.

While Sinn Féin and Alliance leaders will be very happy, the DUP’s leadership has many questions to answer. One is most important to the party itself: the vote assessment. How confident is the DUP that if it forces another election within the next six months that it will do as well then as it has just done?


There will be proposals aplenty to change the existing consociational design, as there have been ever since 1998. Some of these will be fanciful and violate both the letter and the ethos of the Good Friday Agreement. However, if the others had surpassed either the nationalist or the unionist blocs in support (either in votes or seats) there would have been a much stronger case for reconsidering and reforming the rules. The existing cross-community decision-rules weigh the votes of nationalist and unionist MLAs more than those of others.

What is true is that Alliance could conceivably have come second in seats won, but not have been entitled to nominate to the deputy First Minister position – though it would have been entitled to nominate the First Minister had it come first in seats. That problem could be solved by changing the rules so that the two first ministers are also allocated by the d’Hondt rule and by equalising their titles (eliminating the ‘deputy’ designation). Such a rule change could be accompanied by another. The Justice Ministry should be filled in the normal way as other ministries, by the d’Hondt allocation rule. It is anomalous that a nationalist can become First Minister but effectively cannot be the minister in charge of the administration of justice. What clearly will continue to command reformist attention is the unilateral legal ability of the leading party of either major bloc to stop government formation.


Note: The authors are investigators on the ESRC funded Northern Ireland Assembly Election Study 2022.

About the Authors

John Garry is Professor of Political Behaviour at Queen’s University Belfast and leads The Democracy Unit at Queen’s University Belfast.

Brendan O’Leary is Lauder Professor of Political Science, University of Pennsylvania, Honorary Professor of Political Science, Queen’s University Belfast, and Fulbright Fellow to Galway University 2021-22.

James Pow is Lecturer in the School of History, Anthropology, Philosophy and Politics, Queen’s University Belfast.