By Peter Lonergan and Pedro Borges de Carvalho, LL.M. in International and European Public Law (KU Leuven)

While the greatest legacy of the Good Friday Agreement (also known as the Belfast Agreement) has been the end to violence on the island of Ireland, 25 years after its entry into force its implementation has been a mixed success. The power sharing aspect of the agreement has all but collapsed following Brexit, while many of the collaborative bodies created for North-South or UK-Irish relations have been poorly implemented or no longer function.

One of the key issues which has led to this impasse in Northern Ireland is its constitutional uncertainty. On the one hand, this uncertainty is part of the nature of the Good Friday Agreement, which allows the people of Northern Ireland to choose British and/or Irish citizenship, and creates roles for Ireland and the UK in its governance. On the other hand, uncertainty  has been worsened by the negative effects of Brexit on politics in Northern Ireland. This post argues that increasing the clarity of one of the most consequential provisions in the Agreement, the unity clause, might mitigate this uncertainty.

The Good Friday Agreement establishes an obligation for a referendum to be held on a United Ireland if it appears likely that a majority would vote in favour. Specifically, Schedule 1 Paragraph 2 of Article 1 states that:

“The Secretary of State [the UK Minister responsible for Northern Ireland] shall exercise the power under paragraph 1 [to call a referendum, or “border poll”] if at any time it appears likely to him 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.”

Some aspects of this procedure are clear. “Shall” means that the Secretary must call a border poll if they believe that a majority would vote in favour of a united Ireland. “A majority of those voting” indicates that it is not necessary for a majority of the people in Northern Ireland to be in favour of a united Ireland for a poll to take place (in other words, an absolute majority of at least 50% + one of the population), but only that a simple majority of those taking part in a poll would likely vote in favour.

How to assess whether such a majority exists is not defined in the Agreement, but potential means of assessment include independent polls on the question, which are already regularly undertaken, or political election results.

“Likely” is more ambiguous. Studies have shown that personal definitions of “likely” as a percentage probability range from just above 50%, to as high as a 66% chance of an event happening. As no such percentage probability is defined in the Good Friday Agreement, we can logically conclude that a vote would need to be called if the likelihood of a successful poll lay within this ambiguous probability range.

This logical conclusion points towards the greatest issue in Article 1. How the likelihood of a successful border poll should be assessed and the meaning of “likely” are left completely at the discretion of the Secretary of State for Northern Ireland, who is obliged to call a border poll only “if it appears likely to him” that a majority would vote in favour. This unchecked discretion is completely at odds with the idea of the Secretary of State as a neutral arbitrator who is obliged to call a referendum if conditions objectively require it. It also creates a considerable risk of the Secretary acting in bad faith and not holding a poll when it appears likely from all existing factors that a majority would vote in favour.

The consequences of a Secretary of State acting in bad faith are not limited to the decision not to call a border poll. While the first scenario to come to mind is of the UK government refusing to hold a border poll when a clear majority is in favour of unification, the Secretary could also knowingly call a border poll when it appears unlikely that it would pass in order to delay another poll for seven years (Schedule 1 paragraph 3 clarifies that a poll can only be held once every seven years). This delaying tactic is possible, given the gradual demographic change that has taken place in the region over the past century, as the Unionist political majority in elections has declined from 35% in the region’s first election in 1921, to 5% following the Good Friday Agreement in 1998, to a historic low of 0.5% following the most recent elections in 2022.

Unionists could similarly fear a UK government, which was more open to a United Ireland, prematurely calling a border poll for political reasons. While this decision might appear implausible, the precedent of the Conservative Party including a Brexit referendum as a political promise in 2015 is difficult to forget. The Labour Party has been open to Irish unity in the past, including it as a party priority during the 1980’s, while a political agreement to give the Irish home rule by the UK Liberal party in 1912 almost led to civil war in Ireland, eventually triggering the Irish War of Independence. The uncertainty which this politicization of the unity clause would create for the Unionist Community shows that this provision can be abused to the detriment of all communities in Northern Ireland.

With this in mind, it is clearly in the interest of all parties to the Good Friday Agreement to increase transparency over how the likelihood of a border poll would be assessed and on the decision of the Secretary of State. Furthermore, an analysis of the law of treaties shows that the UK is obliged under international law, as a matter of good faith, to provide this information to other parties and guarantors of the Good Friday Agreement.

Interpreting the Good Friday Agreement in Good Faith pursuant to the Law of Treaties

“We pledge that we will, in good faith, work to ensure the success of each and every one of the arrangements to be established under this agreement.”

(All Parties, the fifth preamble of the 1998 Agreement)

Any interpretative exercise of the Good Friday Agreement must observe the general rule of treaty interpretation embedded in Article 31 of the Vienna Convention on the Law of Treaties (VCLT), which establishes at its core that “a treaty shall be interpreted in good faith in accordance with the ordinary meaning to be given to the terms of the treaty in their context and in the light of its object and purpose.

Good faith is of central importance in treaty interpretation. The International Court of Justice (ICJ) found in North Sea Continental Shelf (para. 85) that good faith created concrete rules. The effect of this principle is therefore to create procedural requirements for states (Kolb, Good Faith in International Law, p. 195). For example, in Gabčíkovo-Nagymaros (para. 142) the ICJ clarified that the rules in the North Sea case established that Parties must attempt to find an agreed solution to disputes within the context of the treaty, and to apply the treaty in a reasonable way to realise this purpose. The ICJ added in Application of the Interim Accord (para. 132) that any negotiations or consultations required by the treaty must be taken with a view to finding an agreement, as good faith excludes the right of states to claim a formalistic, but ineffectual compliance with negotiation obligations, and that states could not simply insist on their own position without contemplating any modification to it. Indeed, states should “pay reasonable regard to the interests of the other” when negotiating as part of an agreement (Fisheries Jurisdiction (Germany v Iceland), para. 202).

Outside of disputes, the ICJ also found in Gabčíkovo-Nagymaros (para. 142) that good faith entails that the purpose and intentions of the parties to a treaty should prevail over any literal interpretation, while the International Law Commission noted in its Commentary to Article 26 of the VCLT that good faith obliges states to not take actions or omissions which are intended to frustrate those purposes and intentions. Obligations of good faith are hence deemed to exist separately to any individual article in a treaty, meaning a state can be found to have acted in bad faith without having breached a specific provision of the agreement, as transpired in Nicaragua v United States (para. 270).

In short, good faith requires states to exercise provisions in the agreement which are under their control in a genuine manner, which considers the interests of other parties, and to meaningfully communicate while implementing their procedural obligations.

In applying this principle to the unity clause of the Good Friday Agreement, it is clear that the Secretary of State should now provide clarity on how they intend to assess whether a majority would vote for a United Ireland in a border poll and communicate their findings to the other parties of the Good Friday Agreement. While this might not have been necessary in 1998 , when there was a clear Unionist majority, the increased uncertainty caused by Brexit, and the gradual but unmistakable move towards a Northern Ireland without a Unionist majority makes it increasingly likely that debates on this matter will arise in the coming years. Therefore, while the Good Friday Agreement did not explicitly call for such guidelines to be set out by the Secretary, the spirit of the agreement makes their total discretion in deciding if conditions exist for a border poll unsustainable. This is particularly the case after several unfortunate years in UK-Irish-Northern-Irish relations since Brexit, which drastically reduced trust in the UK government in Northern Ireland.

If the Secretary of State is nevertheless not obliged to provide clarity on their decision making process at the present time, it should be obliged to when any of the other parties to the agreement, such as the Irish government, or the political parties in Northern Ireland believe that the Secretary should call a border poll. The ICJ case law cited above makes it clear that parties to international treaties like the Good Friday Agreement must attempt to find an agreed solution to any disputes within the agreement itself, and to apply the agreement in a reasonable way to resolve them.

Furthermore, its findings that states must apply their procedural obligations in an effective manner, and to not insist on their own position without contemplating any modification to it, implies that the UK would be obliged to consider and respond to claims by other treaty partners that it should have found that conditions exist for a border poll to be called. The most obvious way to prove that the UK will undertake a genuine analysis would be to publish the guidelines which will be used by the Secretary of State to decide if a successful border poll is “likely”.

Even leaving aside matters such as treaty interpretation however, the UK has a political and moral obligation to provide clarity to Northern Irish citizens. The Brexit fiasco in Northern Ireland, during which the Conservative government openly took sides in Northern Ireland’s fragile power sharing system and arguably weaponized the region’s history for political purposes, had a dramatic and negative effect on both trust in the UK government and confidence in the institutions formed by the Good Friday Agreement.

Increasing transparency on how it will decide the region’s constitutional future would be a first step in rebuilding that trust.