UK Devolution Policy + Independence/Reunification Desire

Status: Proposal and implementation-history document. Devolved executive offices have shipped, but policy details and open phase notes below are not a current shipped-behavior specification. Branch: rework/political-system-update Depends on: UK FM/Mayor wiring (Phases 1-4 of uk-jp-devolved-executives.md)

Motivation#

UK devolved-executive offices now exist mechanically (Phase 1+) and have real elections + seated FMs (Phases 2-3). But the FM role currently has no policy choices specific to the devolution question itself - the major political axis that defines Scottish / Welsh / NI politics IRL.

This adds:

• A Devolution tab on the FM's office page (SCO/WAL/NIR only)
• A three-option Devolution Policy chosen by the FM
• A per-region Independence Desire metric (Reunification Desire for NIR)
• A turn-driven drift engine driven by FM policy + regional/national mood
• A soft electoral hook nudging vote share toward pro-independence parties when desire is high

Scope locked with user#

Data model#

Per-region metric: independenceDesire#

Stored as a single field on StateMetrics (or a sibling collection if StateMetrics is full):

• 0-100 scale (UI label varies: "Independence" for SCO/WAL, "Reunification" for NIR)
• Defaults to 50 for any state without explicit seeding (neutral)
• Bounded [0, 100] after every drift

For NIR the same numeric field represents Irish Reunification Desire; only the UI label differs. The engine math is identical.

Per-FM-seat policy: devolutionPolicy#

Stored on governorOfficeState (already a per-(countryId, stateId) document):

• "anti" | "pro" | "independence" - three discrete options
• Default "pro" when unset (matches the historical baseline for most FM seats - devolved governance without explicit independence campaign)
• devolutionPolicyChangedAtTurn?: number - last-change turn for cooldown enforcement

In UI, NIR's "independence" option is labeled "Irish Reunification"; SCO/WAL keep "Independence". The stored enum value is the same.

Seed defaults#

independenceDesire per (region, preset)#

devolutionPolicy per (region, preset, seeded FM party)#

Engine: per-turn metric drift#

Each turn processIndependenceDesireDrift applies a delta to the metric for SCO/WAL/NIR. Drivers (additive):

The approval drivers are linear and unbounded: contribution is sign(50 − approval) × floor(|50 − approval|) × 0.001 per turn, with fractional points truncated toward zero (so 48.1 acts as 49).

Total range: roughly -0.15 to +0.17/turn at extremes.

Long-run baseline: mean reversion pulls toward 25, not 50. Without sustained pro-indy drivers, sentiment defaults to the established constitutional settlement (devolved governance without separatist momentum), not a perfectly balanced 50/50 split. The rate (0.003/turn) is intentionally small so this is a gentle floor pull, not a fast unwind.

Pace calibration (illustrative - actual paths depend on starting value):

• Max-positive run (Pro-Indy FM, both approvals at 0%, inflation above 5%): +0.17/turn → 70 turns per 12 points of climb. Climbing from the 25 baseline to ~95 takes ~8.5 game-years (2.1 FM terms), reachable only in deep-crisis conditions.
• Plausible-positive run (Pro-Indy FM, both approvals around 30%, high inflation): +0.11/turn → climbing 25 → 70 takes ~9 game-years (2.3 FM terms) to cross the "possibility" threshold.
• Mid-band run (Pro-Indy FM, neutral approvals, moderate inflation): ~+0.05/turn, but mean reversion (-0.003) means net ~+0.047/turn above 25. From the 25 baseline → 70 takes ~20 game-years - independence stays a marginal force without a crisis.
• Max-suppression run (Anti-Devolution FM, both approvals at 100%, inflation below 2%): -0.15/turn → drags down quickly until mean reversion flips positive below 25 and creates a soft floor.

A sustained pro-independence FM run paired with sub-50 governments is needed for a generation before separation looks plausible - matching the SNP / Plaid IRL arc.

Soft electoral hook#

When resolving general elections in SCO/WAL/NIR, two stacked effects apply:

(a) Soft electoral transfer (vote-share redistribution)#

Pro-indy party gets a bonus, unionist rivals split a matching penalty - total votes preserved. Magnitude: (desire - 50) * 0.001 per pro-indy party (so at desire=100, +5pp bonus and split penalty to rivals). Capped to ±5pp per party.

• SCO: SNP (uk_snp) gets bonus; Conservative penalised
• WAL: Plaid (uk_plaid) gets bonus; Conservative penalised
• NIR: Sinn Féin (uk_sf) gets bonus; DUP / UUP split the penalty

(b) Pro-indy high-desire bonus (additive vote gain)#

Stepped, vote-gain-only multiplier on the pro-indy party's votes when desire ≥ 60. No rival penalty - total votes increase. Applied AFTER the transfer above, multiplicatively per pro-indy candidate.

Bonus is region-locked: SNP only in SCO, Plaid only in WAL, SF only in NIR. Applies to commons / snap_commons / regionalCouncil / governor. Bands are strict - desire must reach the next full 10pp threshold before the bonus steps up (48.1 acts as 49, 69 still gets +1.5%).

UI: Devolution tab#

New tab between Overview and Legislation in GovernorOfficeClient, conditionally rendered only when (countryId === "UK" && state ∈ {SCO, WAL, NIR}).

Layout:

• Metric card at top: current independenceDesire with a sparkline of the last 24 turns (~half a game year)
• Driver breakdown: shows the current per-turn delta with each driver itemised ("Your Pro policy: +0.00/turn", "Regional approval at 42%: +0.05/turn", "Inflation 4.2%: +0.10/turn", etc.) - so the FM understands why the meter is moving
• Policy selector: three radio cards (Anti / Pro / Independence [or Reunification]), each with an LARP blurb. Selecting one costs 2 Office AP and stamps devolutionPolicyChangedAtTurn. Disabled during cooldown.

Phases#

Phase 1 - Data model + types (this PR)#

• StateMetrics: add independenceDesire?: number field
• governorOfficeState: add devolutionPolicy?: "anti" | "pro" | "independence", devolutionPolicyChangedAtTurn?: number
• New constants file lib/constants/devolution.ts for cost / cooldown / drift magnitudes
• Helper functions for region eligibility + label resolution

Phase 2 - Seeding#

• Seed independenceDesire per region per preset (table above)
• Seed devolutionPolicy defaults via seedOfficeStates (when creating a new governorOfficeState row for SCO/WAL/NIR, populate the policy from a small map per preset)

Phase 3 - Turn engine: metric drift#

• processIndependenceDesireDrift(currentTurn) turn phase
• Reads each SCO/WAL/NIR state metrics + governorOfficeState + national approval + inflation; writes new independenceDesire
• Logs per-driver attribution for telemetry / UI

Phase 4 - UI: Devolution tab#

• New DevolutionTab.tsx + selector modal
• New API endpoint POST /api/country/uk/region/[id]/office/devolution-policy
• Wire into GovernorOfficeClient.TABS conditionally

Phase 5 - Soft electoral hook#

• Hook into general election resolution to apply the vote-share nudge during SCO/WAL/NIR races. Capped magnitude to avoid pathological outcomes.

Open questions (deferred)#

• Referendum trigger - when desire is sustained at 60%+ for N turns, should a referendum election spawn? (Out of scope for v1 per the "soft electoral hook" decision.)
• Per-policy LARP blurbs - final copy will live in the DevolutionTab.tsx component; placeholders for now.
• UI label override for NIR - getDevolutionLabel(stateId, policy) returns "Irish Reunification" for NIR + "independence" policy; the rest are stateId-agnostic.