Bosnia and Herzegovina 2026: The Engineered Divorce


Contents
Key Takeaways
The election was not a contest. It was a ratification. Across six simultaneous contests on 4 October — state, entity and cantonal, with 7,028 registered candidates including 2,977 women — Republika Srpska's secessionist establishment improved its position, with Milorad Dodik's SNSD taking 34 of 83 seats in the entity assembly against 12 for the runner-up. It is likely (high confidence) that the SNSD governs the Serb entity through the full 2026–2030 term without recourse to pro-state coalitions, because the opposition's 12-seat bloc was built on an anti-corruption agenda the entity's economic base cannot fund.
The ban is the mandate. Dodik is barred from office until 2031 — a one-year prison sentence and six-year political ban handed down in February 2025 and survived politically, with the Constitutional Court upholding the legal basis of the prosecution as recently as July 2026. He ran the campaign anyway. It is almost certain (high confidence) that he remains the selector of every appointment that matters in Banja Luka through at least 2027, because the elected president, prime minister and assembly majority all derive from his party.
The divorce is procedural, not declaratory. In his post-election address, Dodik pledged to claw back competences in defence, border control, intelligence and indirect taxation, and launched an initiative for a new entity constitution timed for 9 January. The entity assembly had already declaratively repealed laws struck down by the state Constitutional Court. It is likely (moderate confidence — collection gaps on the drafting timeline) that the first substantive competence-transfer bills reach the entity assembly by mid-2027.
Washington has moved to the other side of the ledger. US sanctions on Dodik were lifted in 2025, reportedly at President Trump's behest, and the official described as "the principal driver behind Bosnia policy as it relates to Dodik and Republika Srpska" is awaiting Senate confirmation for a diplomatic post. There is a realistic possibility (moderate confidence — capped by collection gaps on the administration's internal deliberations) of an explicit American accommodation with the Serb entity before end-2027.
Belgrade's rupture removes the patron, not the project. Aleksandar Vučić conceded defeat to a student-led movement in Serbia's 25 October parliamentary election. The secessionist infrastructure inside the entity is financed, armed with narratives and institutionally embedded in ways that do not require Vučić's consent to function — though we may be wrong on the pace at which a post-Vučić Belgrade recalibrates its client relationship. It is unlikely (moderate confidence) that Serbia's new government actively sponsors a unilateral independence move before 2028.
Deterrence is real, thin, and grieving. EUFOR Althea lost four personnel in a helicopter crash at Brčko on 29 October, days before the Security Council renewed its mandate unanimously on 30 October — the first fatalities of the mission's two-decade history landing on the one district both entities claim, and a reminder that the force's presence is physical, dated, and finite. It is highly likely (moderate confidence — consistent with four years of calibrated probing) that the year ahead brings boundary-testing incidents below the threshold of armed confrontation, not above it.
Portfolio-level. Over the next twelve months the investable distinction is between exposures underwritten by Dayton continuity (Brčko corridor logistics, EU pre-accession instruments, convertible-mark deposit systems pegged to the euro) and exposures underwritten by entity-level patronage: RS public procurement, the entity energy sector, and any contract whose counterparty is a Banja Luka institution currently legislating its own secession. The first category is priced on a state that still collects indirect taxes; the second is priced on a party that has promised to stop remitting them. Institutions holding Balkan exposure without splitting the two ledgers are insuring a divorce they have not modelled.
1. The Loyal Election: Six Contests, One Riddle, and the Geography of Ratification
The 2026 general election in Bosnia and Herzegovina was not an upset. It was a census with ballots. The Central Election Commission certified 75 political subjects and 29 coalitions for the 4 October vote (OHR, 30 October 2026), and virtually every seat mathematically available went to the incumbent ethnonational machines. In the Serb entity, Dodik's SNSD took 34 of 83 assembly seats, with the "Secure Srpska" movement at 12 and the SDS at 10 (The Western Balkans monitor, 4 November 2026). Its candidate for entity president, Savo Minić, won by a wider margin than any poll had predicted — regional tallies put him near 52 per cent (moderate confidence — single-source figure) — on a promise of a ten per cent pay rise and a seven per cent pension increase (Heinrich Böll Foundation, 4 November 2026).
The state-level picture rhymes. Denis Bećirović, Darijana Filipović and Željka Cvijanović were sworn into the state presidency on 16 November (N1 Sarajevo, 5 November 2026; Anadolu Agency, 16 November 2026) — two incumbents returning and one HDZ replacement for the pluralist Željko Komšić. In the Federation's 98-seat lower house, the SDA took 31 seats, the SDP 16, the HDZ 15 (The Western Balkans monitor, 4 November 2026). Professor Sead Turčalo, dean of Sarajevo's Faculty of Political Sciences, read the results as change inside ethnic blocs rather than a weakening of the bloc logic itself — Dayton's architecture producing alternation without transformation (ECPS, 10 November 2026). That judgment squares with the result set, and it frames the rest of this Bosnia and Herzegovina geopolitical risk assessment: each community outvoted its dissidents.
Turnout matters more than any headline. It rose by roughly twelve percentage points to just under 48 per cent (moderate confidence — single-source tally, October 2026), and the rise accrued to the sides that promised divorce or cartel, not to the civic lists. A mobilisation surge that rewards the two extremities is not democratic consolidation. It is the electorate pricing in partition — and voting for its managers.

Why does this consolidate rather than dilute the secessionist course? Because the entity's politics are a closed patronage economy, as framed in our Western Balkans 2026 assessment: jobs, pensions and privatisation rents flow through the governing party, and the governing party now owes nothing to anyone who favours the common state. The riddle of the election is why a sanctioned, banned, convicted leader translated a courtroom defeat into an electoral advance. The answer runs through Washington, and it is the next section.
For organisations with political-risk, market-access or franchise exposure across both entities: price the country as two jurisdictions sharing a currency, not one jurisdiction with internal friction. Baseline: SNSD 34/83 entity seats plus allied lists; state presidency composed of two incumbents and one HDZ newcomer; Federation lower house fragmented at 31/16/15 with no reform majority; a 12-point turnout surge accruing to ethnically exclusive lists. Treat any coalition formula that pairs SNSD and HDZ at state level as a hard-currency event for the state property dossier, not a routine government-formation step.
2. The Banned Kingmaker: A Conviction, a Delisting, and a Policy Written in Washington
Banja Luka's most consequential politician holds no office. Dodik stepped down as entity president last year under a six-year ban on holding public office (Deutsche Welle, 1 September 2026), the product of a February 2025 state-court sentence — one year in prison plus the ban — for defying decisions of the High Representative. The state Constitutional Court upheld the constitutionality of the very criminal-code amendments under which he was prosecuted this July (European Commission, 4 November 2026). The legal architecture held. Everything else moved.
On the day after the sentence, the entity assembly passed a "foreign agent" registry for foreign-funded nonprofits (regional press-freedom monitoring, February 2025) — a measure that tells you what the leadership feared next, and what it intends to control first. At a Banja Luka government press conference on 19 October 2026, Dodik publicly challenged the legitimacy of a state-broadcaster journalist asking about the Ugljevik plant, coal procurement and announced electricity price rises, calling the entity broadcaster RTRS "our television" (regional media-freedom trackers, 3 November 2026). The Serb entity is being prepared, deliberately, as a closed informational space — the prerequisite of every managed divorce.
Washington changed first. The US delisted Dodik from sanctions in 2025, reportedly at President Trump's behest (Semafor via OC Media, 11 November 2026). In Senate-confirmation reporting, the nominee for US ambassador to Azerbaijan, Alexander Alden, is described by sources as "the principal driver behind Bosnia policy as it relates to Dodik and Republika Srpska" — a charge his former State Department critics consider, if true, a judgement problem of the first order (Semafor via OC Media, 11 November 2026). Congress has begun to push back. At a 2 November hearing, the ranking member pressed precisely on whether lifting sanctions had "emboldened" the entity leader; the witness answer was unambiguous — premature relief had incentivised "this longstanding malefactor and ally of Russia" to keep destabilising (US House Foreign Affairs Committee, 2 November 2026). Moscow's sponsorship of the secessionist track is not a partisan talking point in that chamber; it is testified record. For the ledger of adversarial framing, note how the other side reads this: in Russian and entity framing, carried faithfully by the state media ecosystems on both, the delisting is proof that the West's "anti-Dayton" case against Banja Luka was always political — cited here as framing to be modelled, not endorsed.
So the ban aged badly in one direction only. Brussels kept the pressure architecture in place, per the assessments in our Russia 2026 report on the Kremlin's Balkan toolkit, while Washington dismantled its half of it — and the entity's electorate drew the obvious conclusion in October. The investigation into how that conclusion was manufactured runs through Section 3.
For organisations with sanctions, compliance or reputational exposure: re-run counterparty screening on the assumption that US and EU lists on this theatre no longer rhyme. Baseline: one-year sentence and six-year ban upheld in July 2026 on the European side; US delisting in 2025 with no reinstatement since; a confirmation fight in Washington that could reverse personnel but not, before 2027, the underlying policy. Assume divergence persists through at least the mandate-renewal cycle.
3. The January 9 Project
Open with the date, because Banja Luka does. In his post-election address, Dodik pledged continuity and named the programme explicitly: restore entity competences in defence, border control, intelligence and indirect taxation, restrict state judicial authority, and adopt a new entity constitution on 9 January — the entity's holiday and, not incidentally, the anniversary of a date the state Constitutional Court has already ruled unconstitutional (OHR, 30 October 2026). A new entity constitution is not a press release. It is the legal vehicle through which a federal unit redesignates itself as a state-in-waiting, converts the entity assembly into a constitutional convention, and dares Sarajevo to choose between enforcing Dayton and demonstrating that it cannot.
The pattern is already rehearsed. October 2025: the entity assembly declaratively repeals laws the state Constitutional Court had already struck down (European Commission, 4 November 2026). February 2025, the day after the sentencing: the foreign-agent registry. July 2026: the same assembly refuses, yet again, its long-overdue obligation to appoint state Constitutional Court judges while the Federation also failed to finalise two appointments before the election window closed (European Commission, 4 November 2026; OHR, 30 October 2026). Step by step, the state's referee is being starved of quorum while the Serb entity builds its substitute organs. This is the engineered divorce in legal notation. None of it fires a shot, and all of it is designed so that the eventual confrontation arrives with the entity already holding every document it needs.
Two things complicate the timetable, and honesty about them beats elegance. First, indirect taxation is the state's revenue spine — the Indirect Taxation Authority redistributes the VAT pool that finances both entities and the Brčko District, and abandoning it costs the northern half more than it costs Sarajevo in the first fiscal year. The pledge is a bargaining chip and a loyalty test as much as a plan, though we may be wrong on the sequencing. Second, the High Representative retains the Bonn Powers — the ability to impose legislation and remove officials — and used the toolkit as recently as the criminal-code amendments that jailed the entity leader. What the Bonn Powers gain in authority they have lost in perceived legitimacy inside the Serb entity, where every imposition is now exhibit A in the "original Dayton" restoration narrative the leadership campaigns on (OHR, 30 October 2026). Deterrence that half the country regards as occupation has a short half-life, per the deterrence-on-credit analysis in our rearmament divide assessment. The monitoring logic, therefore, in place of a conclusion: watch the January 9 reading of the new constitution for its treatment of "state competences in transition", watch the entity assembly's first transfer bills for which competence leads — defence signalling is bluffer currency; indirect taxation is the real declaration of war — and watch the state court's quorum arithmetic, because an entity that finishes hollowing the referee has already decided the venue of the confrontation.
For organisations with contractual counterparties at entity level: insert constitution-renunciation and competence-transfer termination clauses into all new paper, dated against the 9 January drafting milestones rather than annual review. Baseline: a constitutional initiative announced from a position of 34/83 entity seats; a state Constitutional Court deliberately kept under-quorate; the four named competences — defence, borders, intelligence, indirect taxation — constituting the full secessionist checklist per the High Representative's own accounting; Bonn Powers unused since the amendments that triggered the ban. Assume the drafting proceeds inside 2027 and price renegotiation triggers accordingly.
4. The Banja Luka–Mostar Axis: Two Vetoes in Search of a State
A state that two vetoes can each paralyse is not a state with two governments. It is a state with none. The October ballot returned a state presidency on 16 November composed of Denis Bećirović, Darijana Filipović and Željka Cvijanović, two incumbents and one HDZ BiH newcomer replacing the pluralist Željko Komšić (N1 Sarajevo, 5 November 2026), rotating through eight-month chairmanships for four years. Below it, nothing moved. The Central Election Commission only confirmed results on 3 November after complaints and a manual recount, and while deadlines now bind government formation at lower levels, no statutory deadline exists at state level at all (Balkan Insight, 3 November 2026). The veto architecture therefore meets the calendar unarmed.
The Mostar half of the axis consolidated the same way the Banja Luka half did. HDZ BiH tightened its hold on the Croat vote, a result Professor Sead Turčalo of Sarajevo's Faculty of Political Sciences reads as alternation inside the ethnic blocs rather than any weakening of the bloc logic, with Dayton's consociational architecture reproducing the divisions it was designed to manage (ECPS, 10 November 2026). The HDZ's long-articulated ambition of a de facto or de jure Croat third entity has not gone anywhere either; it has simply acquired a state-level partner that also wants the common state loosened. Two parties, two separatist grammars, one shared method: withhold consent until the centre pays. Where the Serb track threatens to leave, the Croat track threatens to stay on hostile terms. That asymmetry is the entire negotiating position of both.
The math compounds it. In the Federation's 98-seat lower house the SDA holds 31 seats, the SDP 16, the HDZ 15, the SBiH 9 (The Western Balkans monitor, 4 November 2026). No reform constellation exists without pairing the two biggest rivals, and the entity-level map mirrors the state-level one: fragmentation all the way down, and a twelve-point turnout surge that rewarded ethnic curation (Balkan Insight, 3 November 2026). Governing coalitions at state level have historically required months to assemble even with strong incentives. It is now a realistic possibility (moderate confidence) that no state Council of Ministers is seated before the second half of 2027, because both entity blocs have stronger incentives to campaign against the centre than to operate within it.

For organisations with banking, retail or cross-entity infrastructure exposure: map decision rights and licensing dependencies by entity, not by state, because the entity is where a functioning executive exists. Baseline: state presidency seated 16 November 2026 with rotating eight-month chairs; no statutory deadline for state-level coalition formation; Federation chamber split 31/16/15/9 with 27 residual seats; both governing blocs campaigning against state-level governance costs. Treat state-level regulatory initiatives announced before a Council of Ministers exists as announcements, not policy.
5. The Belgrade Rupture: A Patron Falls, a Project Loses Its Cover
Client projects do not die with their patron. They inherit his creditors. On 9 September Aleksandar Vučić dissolved Serbia's assembly, claiming any delay "would threaten the survival of democracy", and resigned the presidency days later to run for prime minister on his party's list (Reuters, 25 October 2026). The gambit failed. On 25 October he conceded defeat to the student-led protest movement turned electoral platform, on the highest turnout in more than thirty years (Reuters, 25 October 2026; Al Jazeera, 26 October 2026). The Student List proceeded to win an outright majority in the assembly (FEPS, 28 October 2026), a movement with no clear leader and, awkwardly, no agreed candidate for prime minister (RFE/RL, 26 October 2026). Its first parliamentary sitting was scheduled for 17 November, with a presidential election due by 27 December (Centre for Western Balkans Studies, 16 November 2026; FEPS, 28 October 2026).
Banja Luka noticed before anyone else did. Dodik publicly congratulated the winners, pledging that Republika Srpska and Serbia "will continue forward together, maintaining the closest and best possible relations, with cooperation and mutual support on all regional and international issues" (Balkan Insight, 26 October 2026). Read that as hedging, not conversion. A movement elected on anti-corruption and lustration promises (FEPS, 28 October 2026) has every incentive to put daylight between itself and the region's sanctioned secessionist cause, at least while its European audience is watching. The Vučić era's quiet infrastructure of political, media and security-service liaison across the Drina now sits on the books of a government that ran against exactly that kind of arrangement. Congressional testimony a week earlier had already catalogued Belgrade's dual track, EU candidacy alongside defence ties with Russia and China and promoted propaganda lines (US House Foreign Affairs Committee hearing, 2 November 2026) — a dossier the incoming majority has every political reason to close.
Whether it actually can is another matter. Patronage networks survive their founders when the institutions that carried them do, and the SNSD machine is one of those institutions. Europe's strategic bandwidth for the Balkans, meanwhile, remains consumed by the war on its eastern flank, as assessed in our forgotten-theatre analysis of the Ukraine–Russia war, which is precisely the window in which a divorce project prefers to move. It is likely (moderate confidence) that a Student List government quietly downgrades its security-political liaisons with the Serb entity during 2027 while keeping commercial ties intact, and highly unlikely (moderate confidence) that it actively sponsors any unilateral statehood move before 2028 — the movement's entire mandate is domestic reconstruction, and foreign adventurism is the one expenditure its voters did not authorise.
For organisations with Serbia-linked supply chains, media holdings or entity-level counterparties: audit every cross-border relationship built during the Vučić decade now, not at renewal. Baseline: an outright-majority government in Belgrade in formation since 17 November; a confiscation-and-lustration legislative agenda (FEPS, 28 October 2026); a congratulatory message from Banja Luka that commits Serbia to nothing; and no post-election statement from the Student List on the entity question at time of writing. Assume commercial continuity and political cooling proceed together, and screen state-adjacent contractors on the Serbian side for the lustration exposure their owners just acquired.
6. The Vacuum-Keepers: EUFOR's Clock, the Bonn Powers, and Deterrence on Credit
EUFOR Althea does not defend the state. It schedules the state's confrontations. The mission's legal floor runs through UN Security Council Resolution 2795 of 31 October 2025, which renewed both the EU force's authorisation and NATO's Sarajevo headquarters for twelve months, to 31 October 2027 (UN Security Council, 2026). The command machinery is borrowed alliance architecture under Berlin Plus, with the operation commanded by NATO's vice-chief at SHAPE and an EU operational headquarters at Mons (NATO, 29 October 2026). Hungary, meanwhile, announced on 4 November a doubling of its over-the-horizon standby reserve for the mission from 150 to 300 personnel — reinforcement on paper, held on Hungarian soil until conditions require it (Adria Defense, 5 November 2026). Budapest has held the mission's deputy command since March 2026, and its posture toward the secessionist leadership, assessed in our Hungary 2026 report, is the least alarmed of any contributing capital.

The mission's year has already been marked in blood rather than doctrine. On 29 October, an Italian Army CH-47F Chinook operating for EUFOR crashed in the Brčko region, killing at least four personnel in the first fatal accident of the operation's two-decade deployment (Euronews, 29 October 2026). The symbolism writes itself: the loss occurred in the corridor district both entities straddle, six days before the Security Council renewed the mandate unanimously through October 2027 (EUFOR, 30 October 2026). The roughly 1,200-strong mission (Euronews, 29 October 2026) now begins its renewed tour without the accident's causes established and with a public instructed, by the force's own statement, not to spread unverified information — a strange posture for a deterrent, and an informative one.
The civilian half of the deterrence stack is the weaker half. A new High Representative has not been appointed; an acting officeholder has held the Bonn Powers since June 2026 (European Commission, 4 November 2026). The arsenal that jailed the entity's leader — imposed legislation, removal from office — is now in caretaker custody, and the Serb entity knows the difference between a full-power principal and an acting one. Security Council members are saying the quiet part in New York: Denmark and Latvia both told the 30 October session that the entity's leadership is openly attempting to dismantle the Dayton settlement through institutional obstruction, defiance of the Constitutional Court, and seizure of state property (UN Security Council, 10258th meeting, 30 October 2026). Concern is not enforcement. Enforcement is what nobody in the current chain of command is equipped to deliver quickly, and the think-tank literature on both militaries says so: the entity retains a strengthened police outside state command, parades it every 9 January, and formally resolved as long ago as December 2021 to withdraw its consent to state-level defence, while the unified armed forces would risk fracturing along ethnic lines in any severe constitutional crisis (Kosovar Centre for Security Studies, November 2026). The whole deterrent rests on a British-run resolution calendar, a caretaker envoy and an alliance distracted by the southern flank we mapped in our Italy 2026 assessment. It is likely (high confidence — the vacancy has already run six months) that the High Representative position is still unfilled in early 2027, and highly likely (moderate confidence — consistent with four years of calibrated probing) that the year ahead brings boundary-testing incidents — court closures, police redeployments, asset seizures — rather than shooting.
For organisations with personnel, fixed assets or insurance exposure in-country: build monitoring around dated decision points, not ambient anxiety. Baseline: EUFOR and NATO HQ mandates expire 31 October 2027; Hungarian reserve at 300 but out of theatre; four EUFOR personnel killed at Brčko on 29 October with the causes unestablished; acting High Representative since June 2026; entity police parades every 9 January; AFBiH cohesion rated fragile in crisis conditions (QKSS, November 2026). Tripwires: any entity law purporting to transfer defence, border, intelligence or customs authority; any move against state property registers; any coordinated non-remittance by the entity of indirect-tax revenues. Escalate on triggers, review quarterly at most.
7. The Separation Balance Sheet: Coal, Convertible Marks, and the Corridors Around Sarajevo
Divorces are filed in court. They are priced at the till. The entity's separatism runs on an economy that keeps three hostages on its balance sheet. The first is the currency. The convertible mark has been pegged to the euro through a currency board since 1998, one of the few state-level institutions every side obeys (Forbes country profile, 2026). Any secession that breaks the payments union breaks the peg, and the ratings architecture has already priced the country's politics into its paper: Moody's holds the sovereign at B3 with a stable outlook, S&P at B+, both citing a complex political system, moderate indebtedness and continued reliance on international financial institutions (Trading Economics, November 2026; Central Bank of Bosnia and Herzegovina). The second hostage is the tax pool. Value-added tax introduced in 2006 became the state's steadiest revenue line (Forbes, 2026), and the entity leadership's pledge to withdraw from indirect taxation is, on the numbers, a pledge to defund its own public payroll first — as we noted in Section 3, the pledge functions as leverage and loyalty test more than as plan.
The third hostage is the corridor map, and here the engineering is already visible. Brussels ended the earmarking of EU grant funds for one of two suspended railway projects on Corridor Vc because the Reform Agenda decisions that conditioned them never came (European Commission, 4 November 2026). At the same time, the entity's railways agreed a direct international passenger service from Zvornik Novi to Belgrade with Serbia's operator, plus a Croatia connection transiting Volinja, entering the timetable on 13 December 2026 (Transport Community, 6 November 2026). The pattern needs no decoding. Infrastructure that runs through the state gets suspended; infrastructure that runs around the state gets built. Regional growth, per the EBRD's latest outlook, is expected at roughly 3.4 per cent for the Western Balkans in 2026 (EBRD, 2026), enough to keep the lights on, never enough to make divorce expensive. Foreign investment dropped sharply after 2007 and has stayed low, while the informal economy carries whatever the formal one cannot (Forbes, 2026). Coal still powers the entity's grid, and its flagship plant's procurement has just become a forbidden question for journalists (regional press-freedom monitoring, November 2026) — a price-rise story the government would rather tell on its own channel. The energy exposure threading through all of this is mapped in our European energy security assessment. It is likely (moderate confidence) that the currency board and the payments union survive through 2027 even under significant institutional stress, because every faction's savings, pensions and trade settle in that instrument — the rare convergence where the secessionists' own incentives do the stabilising.
For organisations with deposit-taking, retail, FX or logistics exposure: harden the treasury perimeter before politics completes it. Baseline: BAM-euro currency board since 1998; sovereign at B3/B+ stable; entity corridor projects losing EU earmarks while Belgrade- and Zagreb-facing rail reopens on 13 December 2026; announced electricity price rises in the Serb entity with suppressed press scrutiny. Hold entity-level receivables to shorter tenors than state-level ones, escrow VAT-linked contracts, and treat the reopening of around-the-state corridors as a leading indicator that planners expect the centre to matter less, not more.
8. The Brussels File: Accession Without a Negotiator
The country is not drifting off the EU path. The EU path has been quietly withdrawn from service. In its 2026 enlargement communication Brussels records "increasing disengagement from the EU path" by the governing class, an executive and legislature paralysed by the absence of a functioning coalition, and reforms that exist on paper and nowhere else (European Commission, 4 November 2026). The state has yet to appoint an operational chief negotiator since negotiations formally opened (European Commission, 4 November 2026). The administrative floor beneath the file is thinning: 27 per cent of senior civil-service posts at state level are vacant, and 33 per cent in the entity (European Commission country report, 4 November 2026). And the machinery that would make membership real is running in reverse — the value of confiscated criminal assets fell from €8.4 million in 2023 to €6 million in 2025 (European Commission, 2026), even as the entity assembly shields the rule-of-law acquis from applying north of the inter-entity line.

Public opinion, ironically, has not been consulted on the retreat. Support for EU membership has stayed strong even as faith in government collapsed (bne IntelliNews, 25 August 2026), and a working arrangement with Eurojust was signed in October 2026, whose value depends entirely on prosecutors being allowed to use it (European Commission, October 2026). The file's real prisoner, though, is in Berlin and the member-state capitals whose enlargement bandwidth is contracting, per the coalition arithmetic we assessed in our Germany 2026 report and the broader European strategy frame of our 2026 geopolitical strategy analysis: a Union rearming and rationing attention has little surplus conditionality to spend on a candidate that will not spend it on itself. Enlargement leverage works through incentives. Incentives require a counterpart. As of mid-November 2026, there is no counterpart, no negotiator, and no coalition to appoint one. It is highly unlikely (moderate confidence — capped by the unresolved coalition arithmetic) that any negotiating cluster substantively opens before 2028; the Growth Plan instruments will continue to disburse into an institutional void, teaching the entity that obstruction carries no fiscal cost.
For organisations with EU-instrument, standards-compliance or donor-funded exposure: index contract expectations to entity behaviour, not state milestones. Baseline: no chief negotiator appointed; 27 per cent state-level and 33 per cent entity senior vacancies; confiscations trending down; EU funds de-earmarked from suspended entity rail projects (European Commission, 4 November 2026); Eurojust cooperation live on paper since October. Write disbursement-linked KPIs that name the entity and the state separately, and assume audit trails will be tested by the entity's non-recognition of the courts that would enforce them.
9. Bosnia and Herzegovina 2026 Geopolitical Risk Assessment — Three Scenarios
Source: CES Intelligence calibrated judgment, 17 November 2026 — triggers and bidirectional justification below.
Scenario A — The Slow Drain: Managed Dissolution on Paper (~45–50%)
The procedural divorce grinds forward with no dress rehearsal turning hot. The entity adopts its new constitution on the 9 January track, passes declarative transfer bills touching defence, border control, intelligence and judicial authority, and continues starving the state Constitutional Court of judges — while stopping short of executing the one transfer that would cost it money: the actual withholding of indirect-tax remittances. No state Council of Ministers forms before late 2027; the caretaker protectorate drifts; EUFOR's mandate is renewed on schedule on 31 October 2027 after choreographed Security Council theatre. Washington's accommodation deepens quietly — a visit, a signalling interview, no reinstatement of sanctions — and Belgrade's new government cools its political liaisons while the rail line from Zvornik to Belgrade reopens on 13 December as scheduled. The currency board holds because everyone's deposits ride on it. The country exits 2027 poorer, emptier at state level (27 per cent senior vacancies becoming 30-plus), and two documents closer to formal separation.
Requires: the SNSD's 34-seat bloc to hold through the drafting cycle without a defection-driven snap vote; Washington to keep its delisting policy unchallenged by Congress through the October 2027 mandate renewal; the High Representative post to remain under an acting appointment through mid-2027; and the entity leadership to calculate, correctly, that procedural maximalism costs less than a kinetic test of EUFOR.
This scenario holds unless: the entity executes actual non-remittance of indirect-tax revenues, converting the loyalty test into the state's fiscal collapse weapon (pushing the trajectory toward Scenario B); a caretaker or newly appointed High Representative imposes legislation touching the entity's constitution or police on the Bonn Powers precedent that jailed its leader, triggering a physical refusal of enforcement (pushing the trajectory toward Scenario B); or a Washington–Brussels–Berlin alignment produces an enforcement-and-renegotiation package the entity cannot refuse (pulling the trajectory toward Scenario C).
The probability is elevated by the electoral ratification of October — 34 of 83 entity seats for the secessionist core with no statutory deadline forcing state-level governance — by the US delisting record and the Alden confirmation fight revealing where American policy is actually written, and by the six-month vacancy atop the Bonn Powers. It is reduced only by the Belgrade rupture, which strips the project of its regional patron without changing its funding, and by the entity's own dependence on the VAT pool it threatens to leave.
Scenario B — The Collision: A Dayton Enforcement Crisis (~30–35%)
A trigger converts the paperwork into a confrontation with uniforms. The most probable pathway: the entity's assembly moves from declarative repeal to operative execution — a state property register transferred to entity control, or customs and excise revenues physically withheld at entity level, or an attempted arrest of a state judicial official executing a ruling in Banja Luka territory. Sarajevo or an incoming High Representative invokes enforcement. EUFOR — roughly 1,200-strong in-country, its reserve now doubled but over the horizon, and its morale tested by the Brčko crash that killed four of its own on 29 October — moves to secure state assets or custody of the convicted leader, whose one-year sentence remains executable. The entity police, built deliberately outside state command and paraded each 9 January, is ordered to interpose. The standoff plays out on camera with the force's capacity at its thinnest and its mandate on a twelve-month clock. Whether it resolves in the entity's climb-down, a fracturing of the AFBiH along ethnic lines as the think-tank record warns, or a spasm of scattered violence in Brčko District — the corridor artery both entities straddle — the settlement itself does not survive the year unchanged.
Requires: the entity leadership to exhaust the patience of at least two of Washington, Brussels and the Peace Implementation Council simultaneously; a conviction that the US accommodation umbrella extends to refusing enforcement, not just refusing sanctions; an operational step by the entity that is unambiguous enough to compel a response but timed so the High Representative vacancy complicates it; and elite cohesion within the SNSD that a conviction-addicted patronage machine cannot easily guarantee.
This scenario holds unless: the entity calibrates forever below the execution threshold, keeping every law declarative and every transfer on paper (freezing the file inside Scenario A); Washington signals, through Senate rejection of the Alden nomination or a formal statement, that the delisting was not a green light — collapsing the entity's assumption of cover (freezing the file inside Scenario A); or the collision itself becomes the forcing event for an internationally brokered constitutional renegotiation (pulling the trajectory toward Scenario C).
The probability is elevated by the demonstrated pattern — declarative repeals, judge-blocking, police parades on an anniversary the state court has ruled unconstitutional — and by the built-in asymmetry: the entity has rehearsed this confrontation institutionally while the enforcing side rehearses it bureaucratically. It is reduced only by the entity's fiscal hostage relationship with the tax pool, by Hungary's doubled reserve signalling allied seriousness at the margins, and by the simple fact that every participant over 40 remembers what the last rupture cost.
Scenario C — The Renegotiation: Dayton II Under Brokerage (~15–20%)
The divorce stops being unilateral and becomes negotiated. A settlement conference — convened by a Washington administration that has already bet its credibility on knowing the entity leadership, seconded by a Berlin anxious about flank stability, blessed by a London holding the pen as conference host — trades formal recognition of entity prerogatives for the survival of a thinner common state: a confederal loosing of competences, an agreed border-regime for Brčko District, state property divided by treaty rather than seized, and the currency board retained as the last shared organ. Belgrade's post-Vučić government, whose movement has no stake in the old patronage war, endorses as the region's proof of pro-European credentials. The SNSD signs because its own electorate, polled as strongly pro-EU even while governed by separatists, prefers a ratified divorce to a contested marriage. The state does not dissolve. It is resized, renamed in function, and thereby survives.
Requires: Washington to convert its accommodation from a delisting into a policy with an architecture — a named envoy, a proposal, a venue; a European trio willing to underwrite a renegotiation that Serbia's new government and the Federation's SDA both accept; the SNSD's leadership to accept a settlement in which the entity wins substance but loses the perpetual emergency it lives on; and a state-level counterpart finally existing to initial the document.
This scenario holds unless: the entity reads brokerage as capitulation and escalates unilaterally rather than negotiate, collapsing the diplomatic track (pushing the trajectory toward Scenario B); Washington's Balkans attention collapses in an election cycle or a redirected crisis, leaving no broker (freezing the file inside Scenario A); or the Federation's parties refuse to legitimise any renegotiation of a settlement they did not lose at the ballot box, voiding the counterparty requirement (freezing the file inside Scenario A).
The probability is elevated by the American policy trail — delisting, the Alden record, a Senate fight that has already put Bosnia on the administration's radar — and by the rare alignment of a chastened Belgrade, a pressured Berlin and an entity that prefers treaties to trials. It is reduced by the absence of anyone to negotiate with on the Bosniak side before late 2027, by the EU's bandwidth rationing, and by the observation that every internationally sponsored Balkan renegotiation since 1995 has been opened by a crisis, never by foresight.

10. Implications
Sovereign and financial exposure. Split the country book into two ledgers now, not at the event: state-level paper (the EUR 800 million 2031 bond at 5.25 per cent yield, B3/B+ sovereign ratings) and entity-level exposure. Cap entity-level counterparty tenors at 12 months through 2027 and stress all BAM positions to a regime in which the currency board survives but the tax pool does not. Escrow every VAT-linked contract at state level, and model the indirect-tax redistribution pool freezing for 90 days as the base stress case, not the tail.
Corporate and operational exposure. Audit every cross-border dependency by route, not by counterparty: assume Corridor Vc-linked state funding stays frozen while entity-facing rail toward Belgrade and Zagreb reopens from 13 December 2026. Redo supplier contracts on the Brčko corridor with dual-routing clauses, and hold 60 days of critical inventory on the Posavina axis. Treat any state-level regulatory announcement made before a Council of Ministers exists as non-binding.
Security and duty-of-care exposure. Build movement and evacuation protocols around four dated windows, not ambient risk: the 9 January entity constitution cycle, any executed entity arrest of a state judicial official, the detention and transfer phases of the convicted leader's one-year sentence, and the 31 October 2027 EUFOR mandate renewal. Pre-position crisis communications for a 48-hour border-friction scenario at entity boundary checkpoints; hold evacuation triggers at 72 hours from the first interposition incident between entity police and any uniformed external force.
Political-risk and sanctions exposure. Run US and EU sanctions lists as divergent datasets through 2027: screen on the union of both, price on the stricter one, and review at each of the three American decision points (the nomination fight, any formal policy statement on the entity, the mandate-renewal diplomacy of autumn 2027). Insert competence-transfer termination clauses — keyed on the four named competences of defence, borders, intelligence and indirect taxation — into all new entity-level paper, dated against the 9 January drafting milestones.
Monitoring and portfolio posture. Stand up a standing watch on five indicators with quarterly re-dating: (1) first operative, non-declarative transfer bill in the entity assembly; (2) the quorum arithmetic of the state Constitutional Court; (3) the appointment or non-appointment of a full High Representative; (4) actual non-remittance in the indirect-tax pool; (5) the seating of a state Council of Ministers. Assign each indicator a scenario direction before it fires: indicators 1, 2 and 4 push toward B; indicator 3 pushes toward A or C depending on the appointee; indicator 5 pulls toward C. Convert the watch to 30-day tempo the day the first operative transfer bill is published.
11. Core Analytical Judgment
The variables in this file are coupled, and the coupling is the finding. An entity that has captured its electorate cannot be dislodged by an election; a state that cannot seat a government cannot be defended by a coalition; a patron that has fallen cannot restrain a client it can no longer pay. Each actor's next move is a function of another actor's absence, which is why the system oscillates: every Bonn Power used manufactures the grievance it was designed to prevent, every sanction lifted purchases calm that must be purchased again at a higher price, every avoided confrontation lowers the cost of the next one. There is no stable equilibrium between the two available models — a state that enforces Dayton and an entity that leaves it — and the October election removed the only instrument (alternation) that could have produced one. The system will therefore not settle; it will ratchet, in the direction of whichever side finds a counterparty. As of this writing, the entity has one (Washington) and the state has none, which is the entire forecast in one sentence. Everything else in this assessment is the mechanics of a divorce that is being engineered precisely because nobody is left to contest it. The state of Bosnia and Herzegovina is not collapsing. It is being archived: carefully, legally, and with the ink of its own institutions.
---
Want deeper analysis?
Confidential advisory programs scoped to your organization — continuous monitoring, dedicated analysis, board-level support. Engagements are annual.
---
DISCLAIMER — This analysis is provided for informational and strategic planning purposes only. It is not investment advice, financial advice, or legal advice, and it should not be treated as such. Probability assessments reflect the analyst's calibrated judgment based on available open-source intelligence as of the date of publication, and are subject to revision as new information emerges. Some quantitative estimates and reported events are based on regional sourcing that may evolve as additional confirmation becomes available.



