How many of the negotiating chapters required to join the EU will Montenegro have closed at the end of 2026?

TL;DR

My central forecast is 30 chapters, with an expected value of 29.24 and an 80% interval of 26–32. Montenegro has already closed 18 chapters, and six of the 15 still open had public draft EU common positions by 10 September (Council status; Chapters 12, 18 and 19; Chapter 1; Chapter 17; Chapter 15). All 33 close in 9% of cases because Croatia and the final rule-of-law and environmental files still have to clear before 31 December (Croatian MFA; Chapter 27 status).

Context

The client’s March baseline is stale. The latest official EU count is 18 provisionally closed chapters after Chapters 8 and 29 closed on 14 July 2026; the 15 open chapters are 1, 9, 12, 14, 15, 16, 17, 18, 19, 22, 23, 24, 27, 31 and 33 (Council of the EU).

A chapter counts only after all member states agree and an EU–Montenegro Accession Conference formally closes it. Ireland can prepare and schedule conferences, but it cannot override a national hold; the Irish presidency itself explains that every member state must agree (Irish presidency).

Evidence

The historical record has two parts. Montenegro spent years at three closed chapters, then moved into a sharp endgame after clearing the rule-of-law interim benchmarks in June 2024 (Council conference). The full closure history is:

Croatia is the closest completed reference class. It had 28 of 35 chapters closed near the end of 2010, then finished the last seven during the next six months; four closed together on 30 June 2011 (Croatian MFA; European Commission). Montenegro needs 15 closures in about three and a half months. That is a much steeper sprint, even allowing for bundled conferences.

The main reason not to extrapolate the old pace is the Council document trail. Draft EU common positions were public for Chapters 18 and 19 on 31 July, Chapter 12 on 4 August, Chapter 1 on 2 September, Chapter 17 on 4 September and Chapter 15 on 10 September (Council records; Chapter 1 record; Chapter 17 record; Chapter 15 record). Recent public draft-to-closure lags were 32 days for Chapter 21, 77 for Chapter 2, 38 for Chapter 28, and 43 each for Chapters 8 and 29 (Chapter 21 draft; Chapters 2 and 28 drafts; Chapters 8 and 29 drafts). The current six therefore have enough calendar time to close if member-state review is smooth.

Chapter 14 is a seventh advanced file, but the evidence is mixed. Reporting says Croatia withheld consent in July over cabotage and related transport concerns, while Montenegro’s European affairs minister later said the chapter was not formally blocked and blamed timing around the summer break (Vijesti; MINA). I treat it as likely to close, but more exposed than the six files with fresh public Council drafts.

The broader pipeline also looks real. Montenegro said on 4 September that it had sent negotiating positions for nine chapters to the Commission and planned to send the other six by the end of September; the announcement did not name the nine (Government of Montenegro). Ireland has made completing Montenegro’s negotiations an explicit presidency aim, and the Council is already drafting the accession treaty in parallel (Taoiseach’s 7 July address; Council status page). The Commission’s 30 June financial package also moves Chapter 33 forward, though it still needs negotiation and a Council common position (European Commission).

The upside stops at the hard files. Montenegro’s working-group leaders said in August that more than 90% of Chapter 24 commitments were complete, but also said remaining laws were due by the end of September and implementation would be watched through December (Vijesti). The government adopted EU visa alignment effective 1 November and Parliament passed constitutional changes on 24 August, both useful steps for Chapters 24 and 23 (visa policy; constitutional amendments). Chapter 27 was still described in April as requiring work on Natura 2000, waste, water, climate administration and inspections (Government of Montenegro).

Croatia is the main political tail risk. It refused to clear Chapter 31 in December 2024, and on 22 July 2026 its foreign ministry tied completion of the talks to a broad list of bilateral issues, including wartime detainees, missing persons, war crimes, property claims, Morinj, the Kotor pool name, the maritime boundary and the training ship Jadran (Plenković’s December 2024 explanation; Croatian MFA). A 1 September meeting with Commissioner Marta Kos restored a constructive channel but announced no settlement (Government of Montenegro).

I modeled the 15 open chapters under four correlated regimes: coordinated completion at 25%, strong but incomplete progress at 45.0%, procedural slippage at 25%, and material disruption at 5%. Within each regime, the six public-draft chapters received the highest closure chances; Chapters 9, 16, 22 and 33 formed the middle group; Chapters 23, 24 and 27 were lower; and Chapter 31 was lowest outside a political deal. Exact Poisson-binomial convolution gives a mean of 29.24, a median and mode of 30, a 50.0% chance of at least 30, a 22% chance of at least 32, and a 9% chance of 33.

What's non-obvious

The obvious forecasts are 24 and 33. A raw extension of the 2026 closure rate lands in the mid-20s. The government’s target lands at the ceiling. Both miss the strongest signal: EU common-position documents for six chapters already exist, and recent files at that stage have often closed within one to three months. That moves the center to 29–30.

Treaty drafting is also easy to misread. It does not add a single chapter to the score. It does show that the EU is spending staff time on a near-term endgame before all chapters are closed (Council of the EU). I read that as a real positive signal, but not one strong enough to erase Croatia or the implementation tests in Chapters 23, 24 and 27.

Uncertainties

• The public Council pages prove that draft positions exist, but several document texts and member-state comments are restricted. I could not verify whether any hidden reservation remains on the six advanced files (Council working-party records).
• Montenegro did not identify the nine chapter positions it said it had sent to the Commission. A chapter-by-chapter Commission acceptance list would narrow the lower tail (Government statement).
• The Council’s published material had no confirmed 29th Montenegro Accession Conference at the cutoff. Four ordinary General Affairs Council dates remained, but accession conferences can be added separately and quickly (Council calendar; 22 September agenda).
• The biggest missing data are final Commission assessments for Chapters 23, 24 and 27, and the private terms being discussed with Croatia. Those facts decide most of the probability mass between 30 and 33.