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 28 provisionally closed chapters, with an 80% interval of 25–31. I assign 7% to all 33 closing and 26% to at least 30 closing. Member-state consent and the rule-of-law-last sequence are the brakes on a full sweep (negotiating framework, Council methodology).

Context

This forecasts the cumulative stock at 23:59 CET on 31 December 2026, using information available by 11 October 2026 at 02:48 UTC. Existing closures count, including those before the question was asked. The verified starting stock is 18 of 33, not the description’s March baseline: the latest completed conference, on 14 July, closed competition policy and customs union (Council, 14 July 2026, country record reviewed 2 October).

The conference scheduled for 12 October remains a future event. Its official agenda, reviewed on 9 October, names Chapter 18, statistics, and Chapter 19, social policy and employment. Those are strong forecasting evidence, but neither is added to the observed baseline before formal closure (Council agenda).

Evidence

The table gives the full closure history, not just the recent acceleration. It contains ten closure-producing conferences, covering 18 December 2012 through 14 July 2026. These are integer administrative counts from official conference announcements and chapter records, as available at the forecast cutoff. The gaps between events contain no additional closures.

The long pause ended after the EU accepted the interim rule-of-law benchmarks in June 2024. That decision unlocked further closures (Council, 16 December 2024). Calculated from the full record above, the subsequent seven closure-producing conferences added fifteen chapters, averaging 2.1 per conference; the largest batch was five. This is a scale check, not a fitted arrival rate. The sample excludes conferences without closures, and the current process is moving faster than the earlier regime.

The strongest upward signal is an actual nine-chapter Council pipeline. The agenda issued on 8 October for the following day’s enlargement working party lists draft positions for Chapters 1, 12, 14, 15, 16, 17, 18, 19 and 22. These are member-state deliberations, not just national targets. Chapter 22 was at presentation and questions stage, while several others had revised drafts (Council, CM 4519/26). Chapter 17 also appears in the preparatory documents for the 12 October ambassadors’ meeting, supporting a high year-end probability even though it is absent from the accession conference’s published closing agenda (Coreper documents).

The Dutch government’s 2 October submission makes that pipeline more concrete. It intends to support Chapters 12, 17, 18 and 19, with continued monitoring of capacity. Its support for Chapters 1 and 15 depends on answers about the timetable for adopting EU law. Chapter 14 was already a proposed closure. I treat those timetable questions as an approval risk, not proof that the chapters will fail (Dutch annotated agenda, attachment 2026D48270).

The files outside that package have different paths. Financial services still required draft amendments to credit-institution and capital-market legislation adopted by the cabinet on 1 October; draft adoption is not enactment or EU approval (government release). Budgetary provisions have stronger institutional preparation: the Commission adopted an accession financial package on 30 June, but expressly left it subject to negotiation and a subsequent common position (Commission release).

Environment is difficult, but it does not require every infrastructure project to be finished before provisional closure. On 1 October, the responsible minister reported a prepared negotiating position and closing report, expected a mid-December closure, and acknowledged implementation obligations continuing through transition periods. That is a credible pathway, not evidence that the EU has accepted the requested arrangements (minister’s interview, reported by Vijesti). The same day’s cabinet release records adoption of waste-management regulations, showing concrete work behind the target (government release).

Croatia remains the main diplomatic constraint. September reporting identifies objections affecting transport, foreign policy and both rule-of-law chapters (Euronews, 24 September). After the 9 October meeting on the training ship Jadran and military-property succession, Croatia’s official statement retained its ownership position and called for further consultations (Croatian foreign ministry). This does not rule out a bargain or an agreed mechanism before year-end. I do not treat it as evidence of a completed bargain.

Domestic rule-of-law work also remains unfinished. The government’s 22 September claim that almost all requirements were met explicitly acknowledged outstanding constitutional changes on immunity and integrity checks (European Affairs Ministry). On 10 October, the governing parties were still seeking opposition support for the integrity amendments and pressing for agreement by 13 October. That is a current execution risk, rather than a reform already delivered (Vijesti, 10 October).

Political support for acceleration is real. The Commission President reiterated the goal of closing the remaining chapters in an account published on 5 October, and the Council established an accession-treaty drafting group in April (EEAS, Council country record). The Council calendar offers later ministerial windows on 19 November and 15 December. Those are scheduling opportunities, not confirmed Montenegro closing conferences (Council calendar entry).

I convert this evidence into an exact scenario-mixture model. All probabilities and scenario averages below are my judgments or calculations, not figures published by the sources. The ordinary branch has these weights:

The shared scenarios create positive correlation: effective coordination helps several chapters together, and a setback delays several together. After the freeze allowance, my marginal year-end probabilities are about 99% each for the scheduled statistics and social-policy chapters, 95% for economic and monetary policy, and 79–88% for the other advanced, comparatively unblocked files. Financial services gets 71%, budgetary provisions 75%, and environment 59.4%. The Croatia-sensitive block has a 19% probability of both transport and foreign policy closing.

The rule-of-law block depends on the other closures. Once every remaining non-rule-of-law file closes, I assign 70% to both rule-of-law chapters closing, 10% to exactly one, and 20% to neither. If other files remain open, only a small exceptional-sequencing allowance permits rule-of-law closure. I also reserve less than 1% for a complete freeze and apply a tiny reopening allowance. Thus the existing stock is not imposed as an absolute legal floor; the Council expressly allows revisiting provisional closures (Council, 14 July). The code supplies every unrounded model parameter.

The resulting mean is 28.0, the median is 28, and the most likely individual outcome is 29. There is 75% probability of finishing between 26 and 30. The central 90% interval is 23–33, wider than the central 80% interval of 25–31.

What's non-obvious

A bilateral hold can cost more than the chapter it directly blocks. Montenegro’s agreed methodology says the two rule-of-law chapters close last; Commission programme material also links their closure to closing the other open chapters (Council methodology, paragraph 10, Commission action document, pages 4–5). Under that normal sequence, keeping transport and foreign policy open also keeps the final rule-of-law pair open. Even closing every other file then produces 29, not 31. This is why technical progress and a finish in the high twenties fit together.

The wording is “should,” not an immutable Treaty prohibition. I retain a small allowance for an agreed departure. Nor does the methodology require the two rule-of-law chapters to close simultaneously; their benchmarks sit in separate common positions (Chapter 23 position, Chapter 24 position). The model favors joint final-package closure, but does not eliminate a split result.

Uncertainties

Public draft-position records do not reveal every member-state reservation or the full negotiated text. I could not verify final EU acceptance of every remaining benchmark, environmental transition arrangement, or diplomatic accommodation. Those are evidence gaps, not proof of failure. The most useful missing evidence would be adopted common positions, a confirmed year-end accession-conference package, the completed constitutional decision, and Croatia’s explicit chapter-closing consent.

Source vintage matters. The Commission’s general country page still displays sixteen closed chapters, while its dated July announcement reports eighteen; I use the dated announcement and the Council’s matching count rather than infer reopening from the lagging summary (Commission overview, Commission, 14 July). The old optimistic article in the question was published on 29 November 2024, so it is not a current readiness assessment (EESC article).

The largest numerical uncertainty is coordinated endgame success, not the starting count. Moving ten percentage points of scenario weight from accelerated progress to strong incomplete progress lowers the mean to 27.8 and the all-33 probability to 4%; moving them the other way raises the mean to 28.3 and completion probability to 10%. The central count is stable. The completion tail is sensitive to diplomatic timing and final parliamentary execution.