The First Yes Is the Crucial Vote

Professor Gauti B. Eggertsson reminds us that referendums have played a role in European integration in more than one context. His broader point is that a popular vote before EU negotiations is not, in itself, alien to European practice. He cites examples from Switzerland, Romania, Moldova, North Macedonia, San Marino, France, Italy and Armenia. However, the Icelandic case deserves to be assessed on its own terms, because not every precedent carries the same weight.

The Swiss example, which is highlighted by Professor Eggertson’s title, should be used with care. The Swiss vote of 2001 was not initiated by the government, but by a popular initiative, “Yes to Europe”. That is an important difference from the Icelandic situation, where the process is politically driven by the government and parliament. In Switzerland, the vote functioned as an outside challenge to the political authorities and asked whether the country should take a first formal step towards EU membership. In Iceland, by contrast, the initiative comes from within the political institutions themselves and concerns the possible reopening of accession negotiations that have already once been started and later suspended. The Swiss case is thus not a close parallel to Iceland.

The remaining examples likewise require careful differentiation. Romania did not hold a referendum on the opening of accession negotiations; those negotiations were already in progress, and the 2003 referendum concerned constitutional amendments that facilitated the country’s eventual accession to the European Union. Moldova’s 2024 referendum was similarly not a simple plebiscite on whether negotiations should commence, as accession talks had already been opened; rather, it sought to give constitutional backing to the country’s European orientation. North Macedonia’s 2018 referendum should also be understood in its specific context: its immediate object was the Prespa Agreement and the consequent change of the state’s name, while EU and NATO integration formed the broader political framework. Armenia presents a different configuration again. Recent Armenian debates have concerned the country’s possible European course, including whether and how such a choice might be democratically or constitutionally anchored, against the background of Armenia’s continuing links to the Eurasian Economic Union, its changing relationship with Russia, and the security consequences of the conflict with Azerbaijan. I had the honour of speaking in Minsk at the fifth anniversary of the Eurasian Economic Community in 2019.

None of these cases therefore constitutes a close precedent for an initial Icelandic referendum on whether to reopen accession negotiations that had previously been commenced and then suspended. They demonstrate that popular votes may arise in the wider context of European integration, but they do not establish a clear analogy for the Icelandic question at hand. San Marino is a special case as a microstate, while France and Italy were established Western European democracies whose referendums concerned different questions. The proper conclusion is therefore that these examples, taken individually or collectively, do not resolve the more important Icelandic issue.

Path Dependency Is the Key Issue

The first referendum of 29 August 2026 would not, as a matter of law, decide accession. A positive vote would give a mandate to reopen negotiations. Membership would still require an accession treaty, approval by the European Union under Article 49 TEU, and ratification in accordance with Iceland’s constitutional requirements. According to the political commitment now being made, it would be followed by a second referendum on the final terms. Without Icelandic internal consent, there would be no accession.

That legal position should be acknowledged. It would also remain true if the parliamentary majority changed before the process was completed. A later EU-sceptical majority could slow the talks, change Iceland’s negotiating position, refuse to submit a final treaty or decline ratification. In strict legal terms, the first vote would therefore not be irreversible. The political reality, however, may be different. Once an accession package existed, the second vote would be held under far greater pressure than the first. The question might no longer be simply whether Iceland should join the EU, but whether it could reject an offer without appearing unreliable or isolated. In that sense, the first vote would be exploratory in law, but it could become the real point of no return in politics.

The promised second referendum is therefore important, but it should not be presented as an absolute safeguard. If it rests primarily on a political commitment rather than on a legally entrenched obligation, a later parliamentary majority could still prevent the process from reaching that vote by withholding legislation, blocking ratification or ending the process. Conversely, if negotiations were completed, the second referendum could take place in a political atmosphere shaped by diplomatic pressure, institutional expectations and appeals not to waste a historic opportunity. For that reason, the decisive issue is not whether a second vote is promised, but whether Icelanders understand the political momentum that a Yes in the first referendum vote of 29 August 2026 would create.

Prof. Dr. jur. Dr. rer. pol. h.c. Carl Baudenbacher, Former President of the EFTA Court, Former Visiting Professor at the University of Iceland