Abstract
© Brian Lawless / Alamy Stock Photo
As the Northern Ireland census records more Catholics than Protestants for the first time, Jon Tonge examines the likelihood of a border poll on a united Ireland.
Northern Ireland’s constitutional future has always been contested. The 1998 Good Friday Agreement emphasised that whether Northern Ireland remains within the UK or becomes part of a united Ireland depends upon the preference of the majority of its citizens. The Agreement declared that a constitutional referendum, more commonly referred to as a ‘border poll’, has to be called if it appears to the Secretary of State for Northern Ireland that there is majority support for a united Ireland. Subsequent border polls may – but need not be – held at intervals of not less than seven years. Recent years have seen an increase in the debate over whether the time has arrived for such a referendum. Sinn Fein, now the largest party within the Northern Ireland Assembly, demands a date be set. This article examines the (sparse) criteria by which a border poll might be called, before assessing some of that possible evidence: opinion polls, election results, unionist and nationalist bloc votes and census results.
The rules of the game
The 1998 Good Friday Agreement (paragraph 1, schedule 2) states that the Secretary of State for Northern Ireland shall call a border poll if at any time it appears likely to him (sic) that a majority of those voting would express a wish that Northern Ireland should cease to be part of the United Kingdom and form part of a united Ireland.
The UK government accepted a ‘binding obligation’ to legislate for a united Ireland in the event of a vote for Irish unification.
No criteria were provided as to how a Secretary of State could judge that a majority might desire constitutional change. At the time this did not appear to matter unduly as a poll appeared so distant. Since the 2016 Brexit vote, however, a referendum has become a more realistic possibility. Sinn Féin listed ‘securing a date from the Irish and British governments for the referenda on unity provided for in the Good Friday Agreement’ as one of its ‘priorities’ in the 2022 Northern Ireland Assembly election. The party topped the poll.
Yet Sinn Fein’s pre-eminence is a necessary but far from sufficient ingredient in forcing a referendum on Irish unity. A Secretary of State requires more extensive evidence that public opinion has shifted in favour of a united Ireland than a one percentage point increase in Sinn Fein’s vote and the party’s supplanting of the Democratic Unionist Party (DUP) as Northern Ireland’s largest. This raises the question of what criteria a Secretary of State could use? Demands for clearer, or published, or fixed (or all three) criteria have increased in recent years and the subject has attracted in-depth academic attention. Legal efforts to elicit fixed criteria have drawn a blank. In the Belfast High Court in June 2018, Mr Justice Girvan declared: The precise circumstances and the political context of a decision are variable and highly political. Decisionmaking in this area requires a political assessment on the part of the Secretary of State and in this case political flexibility and judgement are called for … I am wholly unpersuaded that the Secretary of State is bound to be bound by a policy detailing the way in which that flexible and politically sensitive power is bound to be exercised. In essence it must be for the Secretary of State to decide what matters should be taken into account on the political question of the appropriateness of a poll.
Subsequently, the UK Supreme Court refused to hear a case attempting to force the UK government to publish criteria for a border poll.
The main possible criteria appear obvious: election results and opinion polls – yet they may be problematic. Changes in party fortunes within unionist or nationalist blocs matter less than the total vote for those blocs. Self-enrolling online polls risk manipulation by the politically committed. Face-to-face surveys raise the possibility that there exist shy ‘united Irelanders’ reluctant to admit support. There is no firm evidence for either concern though. Accurate sampling is essential but the annual Northern Ireland Life and Times survey persistently under-reports Sinn Fein’s support. The 2021 version placed the party only level with the Social Democratic and Labour Party (SDLP) even though the latter achieved less than one-third of Sinn Fein’s first preference vote tally at the 2022 Assembly election.
A further complication is that the Good Friday Agreement requires the Secretary of State to take a view that a majority of those voting in a border poll would demand a change to the constitutional status quo. But how do we know who would vote? Average turnout in Northern Ireland elections since the Good Friday Agreement is 64 per cent – yet the Good Friday Agreement referendum produced a much larger turnout of 81 per cent. Self-enrolling online polls reveal (generally impressively accurately) the voting intentions among regular voters. They do not tell us the constitutional preferences of the near one-third of Northern Ireland’s electors who do not vote – some of whom will nonetheless participate in a border poll.
What is public opinion saying?
Figure 1 shows annual polling average support for both main options from the year of the Brexit vote onwards.
‘Poll of Polls’ on constitutional preferences, by year, 2016-22 (%)
It shows the composite results of 30 opinion polls held between 2016 and 2022. Only four showed more support for a United Ireland than for Northern Ireland remaining in the UK. One poll, conducted by OFCOC/ Deltapoll in September 2018, did give a united Ireland overall majority support, at 52 per cent compared to 39 per cent for the status quo. Polls conducted by LucidTalk in December 2017 (48 per cent to 45 per cent) and October 2020 (35 per cent to 34 per cent) and a Lord Ashcroft poll in September 2019 (46 per cent to 45 per cent) indicated a slight favouring for Irish unity over remaining in the UK.
All other surveys have shown more support for Northern Ireland remaining in the UK than for a united Ireland. However, the range has been considerable. The highest figure in favour of the constitutional status quo was 62 percent (Ipsos Mori) in 2016. Support for the Union may never return to such heights. Average support for a united Ireland leapt in 2017 from 25 per cent the previous year to 36 per cent in the immediate aftermath of Brexit but has been steady ever since, never dropping below that mid-30s figure but not rising above 39 per cent. The overall picture from Figure 1 is that support for Northern Ireland remaining in the UK now struggles to command an overall majority of the region’s electors. Yet backing for a united Ireland does not routinely exceed the 40 per cent mark, let alone 50 per cent and has appeared broadly static over the last few years.
A more detailed examination of all polls over the last two years, in Table 1, shows that in all bar two cases, the percentage of ‘don’t knows’ exceeds the gap between the pro-Union versus pro-Unification sides, indicating there may be much to play for.
Opinion Polls on Constitutional Preferences, by Year, 2021-22
Moreover, studies of voters may understate the ‘don’t know’ figure. According to the 2019 Northern Ireland General Election survey, only 7 per cent of voters did not know their constitutional preference but this figure rose to 18 per cent among non-voters.
Unionist versus nationalist support at elections
An obvious alternative to the use of opinion polls as a guide to constitutional preferences is to consider actual votes cast at elections. If the combined vote totals for nationalist parties advocating Irish reunification were to outweigh those cast for unionist parties defending the constitutional status quo, that would substantially increase pressure for a border poll. Figure 2 shows the percentage sizes of the unionist and nationalist bloc votes at Assembly elections since the Good Friday Agreement.
Unionist, Nationalist and ‘Other’ party percentage bloc votes, Northern Ireland Assembly elections 1998-2022
As can be seen, the unionist bloc vote has fallen from a slight overall majority of votes cast in the first contest in 1998, to 40 per cent by 2022. Yet this 10 per cent fall has not been matched by a commensurate increase in the nationalist bloc vote, which has remained virtually static, rising by merely 0.2 per cent across 24 years of elections. The difference between the two blocs is now negligible nonetheless and any further slide in the unionist bloc could, in isolation, offer a case for a border poll. A similar subsiding of the unionist bloc vote has been evident at Westminster elections, from 49 per cent at the last pre-Good Friday Agreement General Election in 1997 to 43 per cent at the most recent in 2019. However, the nationalist bloc vote has also fallen, albeit by less, from 42 per cent to 40 per cent.
The category increasing its vote is that of ‘others’; parties defining themselves as neither unionist nor nationalist, with a near-doubling of their percentage share of the vote in Assembly elections from just over 8 per cent to more than 15 per cent and a similar rise in Westminster contests, from 9 per cent in 1997 to 17 per cent by 2019. A polity designed to create a unionist majority now has three minority traditions: unionism, nationalism and ‘neither of those’. According to the 2021 Northern Ireland Life and Times survey, the ‘neither unionist nor nationalist’ section represents the most common type of elector: 38 per cent, compared to 32 per cent self-identifying as unionist and 26 per cent as nationalist. That Northern Ireland’s election results do not reflect this owes much to the propensity to vote. Unionist and nationalist identifiers were more than twice as likely to vote at the last Westminster election than were ‘neithers’. We know how most unionists and nationalists will vote in a border poll. What will decide the outcome are the swing votes of the ‘neithers’. The party with the strongest support among the ‘neithers’ is Alliance, now the third most popular party in Northern Ireland and currently refusing to take a position on the constitutional question.
The 2021 census results
Recent census results have further fuelled the border poll debate. They showed the original religious basis for Northern Ireland, a contrived two-to-one Protestant to Catholic majority, had expired. In terms of religious community background, the percentage of Catholics, at 46 per cent, now exceeds that of Protestants, by 3 per cent. Catholics also outnumber Protestants in actual religious identification, as Table 2 shows.
Religious identification in Northern Ireland, 2021 census (%)
Although the link between religious affiliation and party voting remains the strongest anywhere in western Europe, mapping such links directly to the outcome of a border poll is problematic. The basis for Northern Ireland’s existence, as confirmed in the Good Friday Agreement is political, the majority constitutional preference of the citizenry, not religious. That Catholics now outnumber Protestants bolsters prospects for a united Ireland, given that a majority support Irish unity (compared to only 7 per cent of Protestants) but neither religion represents an outright majority, as the religious ‘neithers’, like the political ones, have grown.
Identity statistics in the census also increased the optimism of those aspiring to Irish unity. The number of British-only identifiers fell by 116,100 from the 2011 census. At 31.9 per cent, the figure only slightly exceeds the percentage identifying as Irish-only (29.1 per cent) whose numbers rose by almost 100,000. Northern Irish-only identity was stable at 19.8 per cent. Again however, caution needs to be exercised before mapping identity onto a border poll vote, as the hybrid identities of British-Northern Irish, Irish-Northern Irish and British-Irish-Northern Irish all saw increases, as did those of other national identities.
Conclusion
A border poll, let alone a united Ireland, is not imminent based upon current polling evidence. There has been a moderately comfortable lead for supporters of the continuing union of Great Britain and Northern Ireland over those in favour of a united Ireland in every opinion poll conducted in 2021 and 2022. Overall voting levels for nationalist parties are static. That said, there is little cause for complacency among unionists. Opinion polls show there is no longer an overall majority in favour of the Union. Unionist self-identification has declined, as has exclusive Britishness. Yet British unionist decline needs to be accompanied by two things to shift the constitutional dial. The first is Irish nationalist growth and the second is the winning over of the growing centrist bloc of non-unionist and non-nationalist identifiers to support a united Ireland. That will only be achieved by the presentation of an attractive prospectus for Irish unity by its advocates, a task to which they now are turning.
Footnotes
Jon Tonge is a Professor in the Department of Politics, University of Liverpool.