European Social Label
Mbegu Mwongozo 2074
Specification & Criteria: How the Model Runs, and the Basis for Every Number
1. Scope: In, Assumed, Out, and To Develop
A model earns professional trust by naming its own boundaries before anyone else does. Following the Nordic Resilience Framework’s convention, the four registers below state plainly what the model captures, what it assumes openly, what it deliberately leaves out, and what a later iteration would add. The discipline is the one a well-known master’s thesis applied to the sale-of-goods question: rather than litigate whether the law applied, the author assumed openly that it did and then reasoned rigorously from that stated assumption. This model does the same — it states its assumptions out loud and reasons from them, rather than concealing them behind apparent precision.
In Scope
- The seven structural variables (EIS, IX, FDP, RI, FBM, KT, SB/EH) and their couplings (Page 1)
- The behavioural mechanisms and their negative mirrors (Page 2)
- Improvement levers and shock events, each tunable in magnitude, running the three-gate frame
- Five calibrated scenarios: four countries and one regional-integration case
- Trajectory robustness as the output measure — staying on an ascending path under shocks
Assumptions Applied
- The committed thesis — commercial-normative order is causally prior to political order — is assumed, not re-litigated (Page 1 defends it; here it is taken as given)
- Shock transmission channels are fixed by the model; users tune magnitude, not mechanism
- Parameter defaults are point or band estimates from the best available basis, each marked by provenance
- Generational couplings (FDP→SB/EH) operate at assumed lag lengths, marked theoretical
Out of Scope / Not Covered
- Individual psychology below the level of aggregate dispositions
- Political events as such — instability enters only through structural effects on settlement, horizon, and trust; the model makes no attribution about any country’s politics
- Short-run macroeconomic forecasting; exchange-rate and monetary dynamics beyond the rent/fiscal channels
- Climate projection (drought/disease enter as parameterised shocks, not modelled climatologically)
- Any claim to predict; the model is for structural reasoning, not point forecasts
Areas to Develop
- User-editable parameter recalibration (Page 5, first version) toward a fuller engine room later
- The spatial/agronomic map layer (soil, variety suitability, drought gradients) — a separate planned build
- Wider scenario set beyond the initial five; deeper sub-national resolution
- Out-of-sample validation beyond the two retrodiction anchors
- Gendered intra-household bargaining as an explicit modelled layer (currently in the instrument design)
2. The Three-Gate Intervention Frame
Every input — improvement or shock — passes through the same three gates before it affects the trajectory. This is the model’s central operating logic and the reason a well-intentioned intervention can fail or even backfire. The gates correspond, not by accident, to Gauthier’s three compliance conditions (Page 2, §1.1): an action is adopted when it is beneficial, fair, and expected to pay.
Which variables must be past which thresholds for the input to engage at all. A traceability layer does nothing without a registry component in EIS; a rail tax reverses IX if applied below the network-mandatory band. Precondition failure means the input misfires, not merely underperforms.
A coefficient (0–1) multiplying the input’s effect: well-designed and supported, high; rushed or misdesigned, low. Below a floor, the same input with preconditions met turns net-negative — a half-built formalisation that collapses is worse than none. This is the “grey zone.”
Does the input, as structured, reward the behaviour it intends? An input can pass the first two gates and still fail here if it makes honesty more expensive than concealment. The dairy-model case (§5) is the canonical illustration.
One arrow runs backward from all three: failure charges FBM. A botched formalisation, a captured reward, a betrayal of the legible — none is neutral. Each raises formality-betrayal memory and so suppresses the conversion rate on every future input. This is why bad intervention is worse than none, and it is the model’s sharpest message to a programme designer: attempts are not unlimited; each failure poisons the well.
3. Parameters and Their Provenance
Every parameter carries a provenance tag, so a reader knows at a glance how much weight it bears: MEASURED drawn from data or documented natural experiment; ASSESSED from an impact assessment or available data, not independently verified; THEORETICAL from argument or calibration with limited confirmation. Parameters marked Page 5 are user-adjustable within the stated bounds on the parameter-settings page; the rest are fixed defaults shown with their grounding but not handed to the user, on the principle that exposing a parameter a user could only break, not improve, adds noise rather than rigour.
| Parameter | Default | Basis | Provenance | Adjustable |
|---|---|---|---|---|
| Fiscal ignition band (FDP) | 13–15% tax/GDP | Gaspar et al. 2016; confirmed by cross-section — resource-rich states cluster below 15%, highest ratios are non-resource-rich | MEASURED | Page 5 |
| IX network-mandatory band | 35–50% of activity | Inferred from the Kenyan M-Shwari inflection (2012); the point where recorded reputation becomes counterparty-required | MEASURED (single-case) | Page 5 |
| RI suppression threshold | ≈40–50% of revenue | Above this share of revenue from rents, the IX→FDP coupling is regime-changingly suppressed | THEORETICAL | Page 5 |
| FBM decay rate | ≈10%/period | Calibrated; Zimbabwe (maximal, decades-long tail) and Kenya (low) bound the range | THEORETICAL | Page 5 |
| FDP→SB/EH coupling lag | 1–2 generational periods | The committed-thesis arrow; North & Weingast 1689 logic; one supportive observation (Kenya) | THEORETICAL | Page 5 |
| Settlement-breadth bands (SB) | narrow <0.30, broad >0.55 | Khan holding-power typology; V-Dem power-distribution components; EPR dataset | THEORETICAL | Page 5 |
| Elite-horizon bands (EH) | short <5y, long >15y | Olson 1993; leader-tenure and regime-durability proxies | THEORETICAL | Page 5 |
| EIS build rate | +1–2 units/period w/ investment | Calibrated against mobile-money rollout speeds; slow decay | ASSESSED | Page 5 |
| Conflict-gate multiplier vector | dimension-specific | Collier civil-war cost (~a generation); negative (demand-boosting) IX multiplier from Kenya 2008 | MEASURED | fixed |
| Execution-quality floor (Gate 2) | ≈0.35 | Below this, an input with met preconditions turns net-negative; calibrated | THEORETICAL | Page 5 |
The adjustable parameters above are exactly those that meet two tests: adjusting them meaningfully changes the trajectory, and a reasonable analyst might genuinely disagree about the default. Obscure internal constants that no one has a view on, and that moving would only add noise, remain fixed defaults. This curation is deliberate — a settings page that exposes ten consequential, well-explained parameters is more credible than one exposing fifty and trusting the user to know which matter (Page 5 develops this).
4. Improvement Levers
Improvement levers push the trajectory upward; they are the mirror of shocks and run the same three-gate frame. Each is tunable in magnitude within bounds; the channel — which variables it acts on — is fixed by the model. Coefficients drawn from EUSL’s own impact assessment are marked ASSESSED and are explicitly not independently field-verified; they represent assessed effects on available data, offered for testing and adjustment, not established outcomes.
Containerised storage and processing capacity (e.g. ventilated/cooled storage, drying, milling) reduces post-harvest loss and raises the marketable share of production. Default assessed effect: post-harvest-loss reduction toward ~50% and a productivity uplift, drawn from EUSL’s impact assessment on available data for a horticulture corridor — an assessment and scenario, not field-verified outcomes. Gate-1 precondition: a market to absorb the surplus (without it, reduced loss merely depresses price). Adjustable in magnitude Page 5.
Production recording, traceability, and quality verification build the enforcement-infrastructure stock and raise impersonal-exchange share, because a recorded, verifiable history is the legibility that lets strangers transact (Page 2, §1.1). The mechanism is the same one the Kenyan mobile-money credit layer demonstrated MEASURED; the specific corridor coefficients are ASSESSED. Gate-1 precondition: a registry/identity component already above floor.
Safe water and sanitation reduce the diarrhoeal/waterborne disease burden, lowering child morbidity and stunting and breaking the infection–undernutrition cycle — a capability gain in Sen’s sense, raising the substrate on which every other input depends. Assessed effect from EUSL’s impact assessment; not field-verified.
Recorded, identity-verified marketplaces and programmable transaction instruments raise impersonal-exchange share and mitigate kin-tax pressure (by making income legible to the ledger while plausibly private to the claim network — Page 1, §5.4). Grounded in the documented Kenyan trajectory MEASURED. Instrument vocabulary is generic; no product is named. Gate-3 caveat: targeting and privacy design must reach the earner, or the lever fails incentive-compatibility.
Improved or engineered varieties can raise yield and resilience, but the effect is strongly gated: without enforcement infrastructure to protect seed rights and without adoption trust (low FBM), the lever stalls or reverses — farmers rationally decline a technology they cannot secure or do not trust. The model therefore shows this lever helping under the right conditions and failing on distrust under the wrong ones, which is the honest treatment. Validation must be separated from commercial interest (Page 1) or the lever trips Gate 3. Presented as one lever among several, with its caveat, never as a centrepiece.
Tiered skills (basic → vocational → tertiary) raise the capability substrate that lets every other input deploy (Sen), and universities additionally produce the enforcement layer — the certifiers, adjudicators, and agronomists who staff EIS — and the local research that the spatial layer will draw on. A slow variable with a high long-run coefficient: exactly the kind a short-horizon programme underbuilds and the model reveals as undervalued.
Removing tariffs, non-tariff barriers, customs friction, and SPS fragmentation converts relational cross-border trade (workable only within trusted networks) into impersonal trade among strangers across borders — the Greif threshold operating at the level of a bloc. The magnitude is anchored in COMESA’s own figures: intra-COMESA exports stand at roughly US$14 billion (2024) against an estimated potential of about US$101 billion were extra-regional trade diverted inward — the realised level near an eighth of potential, with the gap attributed to weak productive capacity, non-tariff barriers, and slow FTA implementation. The historical calibration point: intra-COMESA trade rose roughly six-fold to about US$19.3 billion in the years after the 2000 FTA. Gate-1 precondition: sufficient productive capacity and enforcement substrate to use the opened market — barrier removal without it strands the gain. This is the scenario that demonstrates the whole thesis at bloc scale.
5. Shock Events
Shocks push the trajectory downward. Each is tunable in magnitude, length, and severity within bounds, but its transmission channel is fixed by the model — a user sets how severe a drought is, not whether drought boosts trust. This constraint is deliberate: free re-routing of a shock’s mechanism would make the model a toy. The pedagogical payoff is that distinct shocks are distinct mechanisms, not intensities of one slider — drought and famine, in particular, are different things.
| Shock | Transmission channel (fixed) | Primary variables | Recovery profile |
|---|---|---|---|
| Drought | Productive-substrate hit; degrades output; can tip into famine if severe | Dim 3; → KT pressure | Recovers with rains unless it tips |
| Famine DISTINCT | Sen entitlement collapse — breakdown of exchange relations through which people command food; not a production number | Dim 3 + capability + conflict-gate probability | Long tail; capability damage persists |
| Epidemic | Capability suppression (labour, human capital) + instability raising gate probability | capability gate + Dim 3 + gate-prob | Medium; depends on health-system EIS |
| Crop disease | Productive substrate with a contagion-then-containment curve and recovery lag | Dim 3 (variety-specific) | Spread, then contained; lagged recovery |
| Oil / commodity price TWO-SIGNED | Exporter: rent windfall raising revenue while suppressing FDP. Importer: cost shock to the substrate | RI (exporter) / Dim 3 cost (importer) | Tracks price path; the Ghana-2011 mechanism |
| Conflict (gate) | Multiplicative gate, dimension-specific multipliers; a flicker can boost IX demand (Kenya 2008) | all dimensions × gate vector | Collier: ~a generation; elevated recurrence |
| Fiscal / rail tax TIMING-SIGNED | Sign flips at the IX threshold: pre-band reverses IX and starves FDP; post-band absorbed | IX, FDP | Immediate; FBM tail if mistimed |
| Formality betrayal | Charges FBM — the state defects specifically against the legible | FBM, IX | Longest tail (Zimbabwe: decades) |
5.1 The governance & security class — neutrally framed
Governance and security disruption enters the model as a small, defined class of structural shocks, named by mechanism and carrying an explicit disclaimer. The model does not name, score, or attribute political events; it processes only their structural effects through variables it already contains.
| Shock | Modelled only as | Existing mechanism used |
|---|---|---|
| Political instability | A shock that narrows settlement breadth and shortens the elite horizon | SB ↓, EH ↓ (Page 1, §7.3) — no new mechanism |
| Predation / expropriation | A shock that charges formality-betrayal memory and feeds rent capture | FBM ↑, RI-path (Zimbabwe land expropriation as documented anchor) |
| Security breakdown | A multiplier on the conflict gate | Conflict gate (§6, Page 1) — already specified |
5.2 The dairy model — the canonical incentive-compatibility case
One worked case anchors Gate 3, the way Ghana’s e-levy anchors rail-tax timing. A reputation or quality system that punishes a single fault makes disclosure existential and so breeds concealment — when one contaminated batch can ruin a whole consignment and liability cascades, the rational actor hides the fault. The documented fix (from a Nordic dairy quality regime): a self-reported fault reduced liability, and testing was free and universal, so the probability of detection approached certainty. Together these made honesty the cheapest path — disclosure dominated concealment in expected value. The model encodes three requirements from this case: graceful failure (accumulated reputation buffers a single fault; decay is conditional on disclosure), asymmetric liability (declared fault costs less than concealed-then-discovered fault, by a wide enough margin), and cheap universal detection (a required EIS component; its absence is a Gate-1 failure that turns the reputation lever negative). The principle generalises: reasonableness is a rational strategy only when the institution is built so that honesty is also the cheapest strategy. The same logic defeats both concealment and bribery — make the honest path the cheapest and detection near-certain.
6. The Five Scenarios
Each scenario is a calibrated starting point against which inputs and shocks are run. Four countries and one region were selected on a single discipline: each must demonstrate something the others do not, and each must rest on strong data. Baseline figures below are drawn from the sources cited; full provenance is on Page 1 and in the research record.
The canonical ascent and a primary retrodiction anchor (Page 1, §8.1). Threshold-straddling fiscal base meets a maturing impersonal-exchange layer; the credit-conversion inflection (M-Shwari) fires the logistic. Demonstrates the positive base case — what engagement looks like.
The cautionary case and second retrodiction anchor (§8.2). Rent inflow arrives while IX is embryonic, suppressing FDP; surface success (fast growth, stable democracy) masks substrate erosion, ending in default and a formality-betrayal event. The only scenario showing healthy headline numbers over a hollowing core.
Strong state capability and very high impersonal-exchange share, but a narrow settlement — the SB/EH tension no other scenario stresses. Tests whether capacity and IX alone secure a trajectory, or whether settlement breadth remains the binding constraint. The deliberate-threshold case, carrying its concentration risk honestly.
The encompassing-interest proof and the counter to Ghana: the same resource-rent problem, the opposite outcome. The kgotla settlement gave elites a long horizon; expenditure was deliberately decoupled from volatile revenue and invested in substrate. Demonstrates that RI is suppressible — that a long elite horizon defeats the curse.
The region as scenario, and the case that speaks most directly to the audience. The lever is barrier removal; the outcome is the trade-and-trust gain when the productive and enforcement substrate can use the opened market — or the stranded friction when it cannot. Realised intra-regional trade near an eighth of potential, the gap attributed to exactly the model’s mechanisms. Paired with Botswana, the two scales of one thesis: a country defeating its curse, a bloc converting fragmentation into surplus. This scenario is fixed in the set; if a build constraint forces a reduction, a country gives way before the region.
7. The Calibration Contract for Pages 4 and 5
This specification is the contract the two interactive pages honour. Page 4 (the simulator) runs everything on the defaults above: the user selects a scenario, adds improvement levers and shocks, tunes their magnitude/length/severity within bounds, sets sequence, and reads the trajectory — with the outcome shown first and components beneath, so the verdict is legible at a glance and the anatomy available beneath it. Shock mechanisms are fixed; only magnitude is tunable. Every live parameter carries a popup giving its default, basis, provenance, and meaning. Page 5 (parameter settings) exposes the curated set of consequential, genuinely-contestable parameters (those marked Page 5 above) for adjustment within bounds, each with a popup explaining what a change does and in which direction; it also allows a criterion or mechanism to be deactivated, at which point the corresponding control on Page 4 greys out. One clean view, the parameters that matter — clarity over completeness, because this is a model, and a legible model is a more useful one than an exhaustive but unreadable one.