SCALE-UP EU
⚠️ CIRCUIT BREAKER ACTIVE
🌊 Complete Value Distribution Waterfall
📈 Visual Distribution (% of Gross)
🎯 Pre-configured Scenarios
📊 Scenario Comparison
EU Grant
€360K
90% of total project
Co-financing needed
€40K
10% legal minimum
Total project
€400K
Eligible cost
Leverage
10x
EU funds / own funds
🔄 Co-financing Flow
📋 Priority Sources

✅ Priority #1 — Technical Assistance (TA)

Municipal staff involved in implementation are paid from EU TA funds (Art.46 Reg.1083/2006). These salaries are legally redistributed as increased compensation. Net effect: municipality receives EU project funds WITHOUT spending own budget. This is the core argument for mayoral partnership.

✅ Priority #2 — Local Social Fund

Up to 30% of Local Social Fund can be allocated as co-financing if the project purpose is social (mobility, accessibility, digital inclusion). Eligible per ESIF 2021–2027 regulations. Does not affect distributable surplus.

⚠️ Priority #3 — Collective ICA Reserve (last resort)

For major strategic projects only. Requires General Assembly vote. Maximum 20% of reserve per project. Protects the cooperative's credit rating for future STO bond issuance (2027–2028).

🏛️ Municipality Stake in the Cooperative
📊 Local Municipal Impact
📊 SROI Calculator
Social Return on Investment (SROI): Banke-Thomas et al. (2015, BMC Public Health, DOI:10.1186/s12889-015-1935-7) found SROI ratios from 1.1:1 to 65:1 across 40 public health interventions. Formula: SROI = Total Adjusted Social Value / Total Investment. Cepiku et al. (2022, Meditari Accountancy Research, DOI:10.1108/medar-05-2021-1307): systematic review confirms SROI as gold standard for cooperative impact measurement.
SROI = (Social Fund + Jobs Value + CO₂ Value + EU Leverage) / Total Investment
€3.20
For every €1 invested, the cooperative generates 3.20 in social value
Reference range: 1.1:1 to 65:1 (Banke-Thomas et al., 2015). Mondragon: estimated SROI ~8:1 based on Basque employment multiplier data.
👥 Distribution by Member Type
⚖️ Salary Scale (6:1 Mondragon)
📊 Patronage Split

🏭 Mondragon Cooperative — Benchmark Comparison

Mondragon Corporation (Basque Country, Spain): world's largest worker cooperative. Founded 1956. 70,000+ workers, 257 companies, €12B+ annual revenue (Wikipedia 2024). Less than 5% bankruptcy rate over 60 years despite Spain's severe financial crises.

Salary Ratio
3:1–9:1
avg 5:1 (vs market 300:1+)
Bankruptcy Rate
<5%
over 60 years of operation
Entry Deposit
€15K
per member (socio)
Cash/ICA Split
30/70
OnTime target: 40/60
Sources: Rodriguez-Oramas et al. (2022) Frontiers in Psychology DOI:10.3389/fpsyg.2022.806442 · Unterrainer et al. (2022) Frontiers in Psychology DOI:10.3389/fpsyg.2022.947559 (77 studies, 83 enterprises — 63.5% resisted degeneration) · Corporate Rebels (2024)
🔮 Plural Stack — State of the Art Governance

RadicalxChange Plural Stack integrated into MobiCoop SCE governance architecture. Each tool maps directly to a structural need identified in the cooperative's founding philosophy (Wolff, Weyl, Buterin).

🗺️ Integration Architecture
Revenue → Plural Money (MobiCoin) per municipality
  ↓ exit tax 5% → Local Social Fund
Local Social Fund → QF rounds (max 30 projects) + QV allocation
IP contributions → PCO licensing (Celindra / FanCourier) + Shapley royalties
Membership → SBT / Human Passport (anti-sybil, Holonym)
Governance votes → QV (nuanced, 1-member-1-vote preserved for constitutional)
External capital → STO bonds (€100 min, MiFID II, 2027+)
🗳️ Quadratic Voting — Social Fund Allocation

Each member receives 100 voice credits per QV round. Cost of votes = votes². Prevents capture by majority while preserving preference intensity. Per Lalley & Weyl (2018), QV achieves utilitarian optimality that 1-person-1-vote cannot.

Credits remaining: 100 / 100
Total Social Fund allocated from QV round: €0
📐 QV Math Explained

Cost Function

1 vote1 credit
2 votes4 credits
3 votes9 credits
5 votes25 credits
10 votes100 credits (all-in)

Why QV for the Social Fund?

• Minority cities with strong preferences aren't outvoted by majority
• Iași can express 10x preference for cycling infrastructure vs Cluj's mild preference for parking
• 1-member-1-vote preserved for constitutional decisions (statute changes, mergers)
• Piloted successfully: Edge City Lanna 2024 (Chiang Mai), 200+ participants
CHI 2025 validated in cooperative co-housing context

📖 Glossary & Science — Research Foundation

30+ key terms from cooperative economics, mechanism design, and governance science. Each entry includes a plain-English definition, the mechanism (how and why it works), and an academic citation. Click any term to expand.

📚 Academic References
🧠 Viable System Model — How the Cooperative Maps to Cybernetic Systems

The Viable System Model (Beer, 1972) identifies five recursive systems every autonomous organization needs. This diagram shows which systems are human-controlled vs AI-controlled in the current target architecture (Phase 3). Click a phase below to update.

S5 Policy / Identity
👤 Human: General Assembly via QV
Sets cooperative values, priorities & constitutional rules
S5 — Policy: The identity-forming system. Beer: “S5 maintains the systemic identity of the viable system.” In a cooperative, this is the General Assembly — the only body that can change the statute, define membership criteria, or override the AI. Under EU law (SCE Reg. 1435/2003 Art.36), this body cannot be delegated to an AI.
🛑
Democratic Boundary
Members retain constitutional authority — humans set the destination, AI flies the plane
S4 Intelligence / Adaptation
🤖 AI Strategy Agent
Monitors market, proposes parameter adaptations
S4 — Intelligence: Looks outward at the environment (market conditions, regulatory changes, competitor signals) and translates into adaptation proposals for S3/S5. AI excels here: real-time data aggregation, scenario modeling, macro signal detection. Ashby’s Law: S4 needs variety ≥ environment’s variety. AI achieves this; humans cannot.
S3 Optimization
🤖 AI Operations Agent
Waterfall optimization, circuit breakers, resource allocation
S3 — Optimization: Manages internal stability and resource allocation across S1 operations. Runs the value distribution waterfall, triggers circuit breakers when volatility exceeds 20%, and resolves resource conflicts. This is exactly what this simulator demonstrates: the AI runs S3 computations instantly across 12+ parameters.
S2 Coordination
🤖 AI Coordination Agent
Harmonizes city operations, resolves inter-municipality conflicts
S2 — Coordination: Prevents oscillation between S1 units (the individual cities). When Iaşi and Cluj compete for the same EU funding window, S2 mediates without needing S3/S5 to intervene. AI coordination agents can run 24/7, resolving scheduling and resource conflicts in milliseconds.
S1 Operations
🤖+👤 AI + Workers
Day-to-day operations in each city (Iaşi, Cluj, Timişoara)
S1 — Operations: The actual mobility services: smart parking, shuttle dispatch, MobiCoin transactions. Workers remain essential for embodied tasks, customer relations, and edge cases. AI handles routing optimization, anomaly detection, and billing. Hybrid: AI augments, workers decide edge cases.
🗳️ Phase Roadmap — Transition to AI-Governed Cooperative

Click a phase to see how the VSM controller assignments change over time. Default: Phase 3 (AI-Delegated) — the current target architecture.

Phase 1
AI-Assisted
Years 1–2
All S1–S5 human-controlled. AI provides dashboards and alerts only. No autonomous decisions.
Phase 2
AI-Augmented
Years 3–4
S1–S3 partially automated (routine operations). S4–S5 remain fully human-controlled.
Phase 3
AI-Delegated
Years 5–7
S1–S4 delegated to AI. S5 remains human (QV general assembly + ethics board).
Target
Phase 4
AI-Autonomous
Year 8+
S1–S4 fully AI with multi-agent coordination. S5 human: constitutional authority only.
⚖️ EU Legal Framework — SCE + EU AI Act Compatibility
💡 In Simple Terms

Under current EU law, a cooperative must have human governance bodies (general assembly, management board, supervisory board). The AI cannot legally replace these — but it can serve as the executive instrument that carries out decisions made by members. Think of it like this: members vote on the destination (via Quadratic Voting), and the AI flies the plane. The EU AI Act requires that humans can always take back control — which is exactly what the emergency override clause guarantees.

⚠️ AI Governance Risks & Safeguards
🔄
Algorithmic Monoculture
A single model flaw or dataset bias cascades identically across all cities in the cooperative ecosystem, amplifying harm rather than containing it.
🛡️ Safeguard: Mandate algorithmic diversity — different model families per city, ensemble voting for high-stakes decisions, scheduled adversarial red-teaming.
🎯
Adversarial Attacks
Prompt injection, data poisoning, or model inversion attacks could manipulate AI resource allocation, redirecting funds or distorting QV outcomes.
🛡️ Safeguard: Formal verification of decision boundaries, anomaly detection on input distributions, cryptographic audit trails for all AI state changes.
📐
Value Drift
Reward misspecification causes the AI to optimize an economic proxy that gradually diverges from cooperative principles — efficiency over equity, growth over democracy.
🛡️ Safeguard: ICA Principles encoded as hard constitutional constraints. Any optimization violating Rochdale principles triggers automatic human escalation via S5.
🚨
Emergency Override
The system must allow full human takeover at any time — EU AI Act Art.14 requires meaningful human oversight for high-risk AI systems. Cooperative democracy demands it.
🛡️ Safeguard: Supermajority QV override available within 24h, freezing all AI decisions. Physical kill switch. EU AI Act-compliant shutdown procedure documented in governance charter.