Afrimintel › Methodology › §10 Bankability and Risk-Adjustment
§10 — Bankability and Risk-Adjustment Methodology
Methodology v1.2 publication · Effective 13 May 2026 (v1.0.55 deploy) · Editorial responsibility: Nikesh Patel · Status: PUBLISHED — the analytical layer that pairs with §07 province scoring to support institutional-grade decision workflows.
What §10 does. §07 Table 4 produces a province-level score that combines mineral system potential (MSP) with institutional context (IC) under an opportunity multiplier (Opp). The province score is the geological substrate — the answer to "how prospective and how operationally tractable is this province at the macro level?" §10 extends this with an asset-specific analytical layer that supports the question institutional capital actually asks: "is this specific deposit bankable under DFI / project finance / streaming counterparty / strategic acquirer evaluation, and what are the downstream risks?" §10 is analytically additive to §07 — it does not modify province scores; it provides asset-level bankability and downstream-risk decomposition that institutional users decompose into investment decisions.
§10.1 — The six bankability dimensions
For any intelligence-grade institutional dossier, six dimensions decompose the bankability question:
| # | Dimension | Substrate | State convention |
|---|
| A | NPV reconciliation against published economic study | Operator NI 43-101 / S-K 1300 / JORC technical report; Afrimintel DCF tool reconciliation | Sourced (published study) + Derived (Afrimintel reconciliation) |
| B | Commodity-market scenario decomposition | Conservative / Base / Critical-minerals-demand scenarios; IEA STEPS/APS/NZE references for battery-mineral assets | Derived (scenario applied to operator base case) |
| C | Demand-elasticity context | Long-term commodity demand growth assumptions; supply-side context (LoM additions; brownfield expansions; new discoveries) | Derived (Afrimintel synthesis from published forecasts) |
| D | Risk-adjusted discount rate framework | Country-risk-adjusted discount rate; v1.0.46 platform convention: 8.0% + (10−IC) × 0.8% | Derived (formula applied to §07 IC) |
| E | Counterparty-eligibility threshold | DFI debt eligibility / streaming-royalty eligibility / strategic-acquirer eligibility per asset profile | Derived (Afrimintel assessment against published DFI mandate criteria) |
| F | Sensitivity decomposition | ±10% / ±20% sensitivity on price, capex, opex, discount rate | Derived (DCF tool output) |
§10.2 — The three downstream-risk dimensions
Paired with the bankability dimensions, three downstream-risk dimensions decompose the institutional risk question:
| # | Dimension | Substrate | State convention |
|---|
| 1 | Community risk | ACLED admin1 conflict-event density (trailing 12 months); resettlement context (IFC PS5 / AfDB OS2); indigenous peoples context (IFC PS7); cultural heritage (IFC PS8) | Sourced (ACLED) + Inferred (PS-specific assessments pending primary-source ESIA) |
| 2 | Environmental risk | Water basin context (TNFD WWF Water Risk Filter); biodiversity priority area (IBAT); ecosystem services (ENCORE); IFC PS3/PS6 alignment | Sourced (TNFD tools where applicable) + Inferred (PS-specific assessments pending primary-source ESIA) |
| 3 | Geopolitical risk | EITI fiscal-terms disclosure; sovereign engagement context (WB CPF / AfDB strategy); contract-terms surface (per EITI 2025 work where applicable). For named-context targeted deepening, see EITI-C Simandou (v1.1.2 publication; December 2025 contract disclosures). | Sourced (EITI/CPF where available) + Inferred (asset-specific fiscal impact pending contract disclosure) |
§10.3 — Multilateral safeguards framework integration
For institutional-grade dossiers, three multilateral safeguards frameworks are materially in scope and assessed in parallel rather than IFC-only. This reflects the Africa-context reality that AfDB ISS and WB ESF have distinct applicability windows that are not subsumed by IFC PS alignment.
IFC PERFORMANCE STANDARDS (PS1-PS8)
The eight IFC Performance Standards form the substrate for any private DFI lending consideration. PS1 (Assessment and Management of E&S Risks) is universally applicable; PS2-PS8 vary by asset profile. Per-asset PS applicability is surfaced in the Component E v1.2 panel on each worked dossier page.
WORLD BANK ENVIRONMENTAL AND SOCIAL FRAMEWORK
The WB ESF (replaced the Safeguard Policies in 2018) applies via country engagement strategy. Country Partnership Framework (CPF) proximity determines material applicability — assets in jurisdictions with active WB extractive-sector engagement carry higher WB ESF proximity than assets in post-coup or sanctions-context jurisdictions.
AfDB INTEGRATED SAFEGUARDS SYSTEM
The AfDB ISS (Operational Safeguards OS1-OS5) is the AfDB-specific safeguards framework. AfDB is a material regional DFI lender for African extractive projects; comparative ranking among regional DFI lenders for specific jurisdictions is asset-context dependent. OS1 (Environmental and Social Assessment) is universally applicable; OS2-OS5 vary by asset profile.
Three-column applicability surface. The Component E v1.2 panel on each worked dossier surfaces IFC PS / WB ESF / AfDB ISS in parallel three-column format with tier badges (MATERIAL / RELEVANT / CONTEXT) and state tags (Sourced / Pending / Inferred). This reflects the Africa-context reality more accurately than IFC-PS-only framing.
§10.4 — TNFD biodiversity substrate
Taskforce on Nature-related Financial Disclosures (TNFD) emerged from 2023 and is increasingly embedded in DFI E&S due diligence. §10 surfaces a per-asset TNFD substrate carrying:
- Water basin context — the river/lake basin the asset operates in (e.g., Kagera River → Lake Victoria basin; Lualaba → Congo Basin upper)
- Biodiversity priority area classification — whether the asset is in an IUCN-protected area, IBA-classified area, or UNESCO-listed area
- Ecosystem services — material ecosystem services the asset interacts with (freshwater regulation, agricultural support, fisheries)
- Applicable TNFD tools — which of the 5-tool TNFD catalogue (WWF Water Risk Filter, IBAT, ENCORE, Aqueduct, Global Forest Watch) materially applies
TNFD outputs are tagged per the Quality Standard three-state discipline: Sourced (basin and protected-area context is geographic, verifiable from published sources), Inferred (specific ecosystem service classification pending site-specific assessment), or Pending (TNFD tool outputs to be derived at panel deploy where the tool itself is applied).
Externality-monetisation companion. Where TNFD identifies nature-related dependencies and impacts qualitatively, the
SAVi-Mining Adaptation v1 (v1.2.0 publication) extends the same substrate into an explicit externality-monetised NPV — applying IISD's published SAVi methodology to a worked African mining asset. Kabanga is the development pilot; the methodology ships in-progress with conventional NPV anchor Sourced, externality category identification Derived, and every monetised externality value Pending the ESIA input layer. Populated SAVi-extended NPV is a v1.2 Phase 2 deliverable. The adaptation is SAVi-aligned per IISD published documentation; no IISD collaboration is confirmed and load-bearing Counterparty Extension language applies.
§10.5 — Discount rate framework (cross-reference v1.0.46 platform convention)
Per v1.0.46 platform convention, the asset-specific country-risk-adjusted discount rate uses the linear approximation:
rasset = 8.0% + (10 − ICprovince) × 0.8%
This anchors at 8.0% for IC=10 jurisdictions (Australia-equivalent institutional context) and rises linearly toward 16.0% as IC approaches 0. The 0.8% per-IC-point coefficient is calibrated to industry-standard country risk premium ranges (~1-8% over reference base) within the IC=1-9 span observed across the 13 African provinces.
Operator-discretion variance ±2pp around the 8-10% benchmark range is observed across an n=3 indicative sample drawn from prior-cycle reference data. Prior cycles cited Ivanhoe Kakula DFS, Lifezone Kabanga FS, and AVZ Manono DFS as the discount-rate reference set; specific values cited in those cycles carry Inferred-propagated status pending primary-source re-verification. Industry practice for AACE Class 3 feasibility studies includes operator selection of discount rate appropriate to capital allocator audience expectations; the operator-discretion variance observation is methodologically grounded independent of the specific n=3 sample values. v1.0.58 downgrade: specific per-study discount-rate values previously cited as concrete (8% / 8% / 10%) are flagged as Inferred-propagated pending re-verification against the cited technical reports.
Asset-level decomposition companion. For institutional users requiring an asset-by-asset decomposition of the discount-rate question — risk-free anchor plus country risk plus commodity volatility plus asset stage plus disputed-tenure and post-settlement-durability premia — see the
Discount Rate Derivation Framework v1 (v1.1.2 publication). The framework is the computational decomposition methodology that sits beside this §10.5 platform convention; the two are not in conflict. §10.5 remains the production discount rate wired into the platform data layer. The framework's coefficient on the IC→π
c conversion (1.0% per IC point) differs from §10.5's 0.8% per IC point; the difference reflects the asset-level sharper differentiation versus the province-level smoother screening, and is named as a v1.3 editorial calibration item in the framework's §3.1.
Policy-perception overlay πp (v1.6.11). The §10.5 base rate derives from the §07 composite IC, which per the v1.6.8 country-band ruling carries a jurisdiction’s
blended Fraser IAI — mineral potential combined with policy perception. For high-mineral-potential / low-policy jurisdictions this systematically under-prices policy risk at the cost of capital (the DRC is the extreme case: IAI 57.46 against a Policy Perception Index of ~18). The bankability discount rate therefore carries an additive policy-perception premium π
p, driven by the Fraser PPI of the asset’s operating jurisdiction (not the blended province IC) and converted per
framework §3.6. π
p is an overlay on the bankability rate only — the platform country-risk
band is unchanged. The PPI→premium tier mapping is a Derived editorial convention pending empirical calibration (v1.3). Applied across the worked dossiers: Kamoa / Manono (DRC PPI ~18, severe, +3.0–4.0%), Loulo (Mali, provisional, Pending), Kabanga (Tanzania PPI 57.61, elevated, +1.0%).
§10.6 — Counterparty eligibility decomposition
The bankability layer surfaces counterparty-eligibility on a per-asset basis. Different asset profiles trigger different counterparty sequences:
| Asset profile | Counterparty sequence |
|---|
| Operating supermajor (Kamoa-Kakula profile) | Corporate balance sheet → bond market → strategic acquirer; DFI debt unusual at this profile |
| Pre-FID greenfield development (Kabanga profile) | DFI debt anchor (AfDB / IFC / US DFC) → commercial bank syndication → equity from strategic counterparty |
| Disputed-tenure (Manono profile) | Pre-resolution: no conventional counterparty eligibility; post-resolution: pathway depends on resolution architecture |
| Post-settlement re-entry (Loulo-Gounkoto profile) | Streaming/royalty (Franco-Nevada / Wheaton precedent) → DFI debt eligibility (durability-trigger gated) |
| Infrastructure / corridor (Lobito profile) | Multilateral cofinancing structure (US DFC + AfDB + Afreximbank + EU GG); equity from logistics consortium |
§10.7 — Province score (§07) vs asset bankability (§10) — analytical relationship
§07 Table 4 produces a province score — a macro-level geological substrate assessment. §10 produces an asset-specific bankability decomposition. The two layers interact in three ways:
- Province IC feeds asset discount rate. The §07 IC composite is the input to §10.5's discount rate formula. Province IC stability is therefore upstream of asset bankability.
- Asset bankability can reveal province IC inconsistency. Where a §10 output reveals that an asset's operational context is materially different from its province's IC composite (e.g., asset operating in a sub-national region with security context different from the IC composite's national average), the audit log records the inconsistency and the editorial decision (resolve at §10 with asset-specific refinement; resolve at §07 with composite revision; flag for v1.3 development).
- Province MSP is upstream of asset prospectivity. Asset-level grade and tonnage data feed back into province MSP calibration over time; this is a Q3 2026 v1.3 prospectivity engine consideration.
§10.8 — v1.2 publication scope and v1.3+ deferrals
v1.2 ships with §10.1-§10.7 codified plus Component E v1.2 panels integrated across 5 worked dossiers. v1.1.2 minor extended the methodology layer with two named additions: Discount Rate Derivation Framework v1 (asset-level six-component decomposition methodology sitting beside §10.5) and EITI-C Simandou (targeted disclosure-data deepening on the Simandou December 2025 contract disclosures). v1.2.0 minor adds SAVi-Mining Adaptation v1 (Component E2 — externality-monetised NPV methodology applying IISD's published SAVi framework, with Kabanga as the development pilot; methodology ships in-progress, populated valuation deferred to v1.2 Phase 2).
v1.3+ deferrals:
- Multi-commodity volatility weighting (cross-reference: Discount Rate Derivation Framework v1 §2.3 — πv commodity-volatility premium) — for assets with material by-product credits, primary-commodity σ is first approximation; v1.3 refines to weighted multi-commodity volatility.
- Calibrated band-to-discount-rate propagation (cross-reference: Discount Rate Derivation Framework v1 §2.2 — πc country-risk premium) — v1.2 uses IC point estimate; CPI-proxy band preserved in source attribution but does not propagate to discount-rate band. v1.3 extends propagation.
- Severity-weighted, proximity-weighted ACLED scoring (§10.2) — v1.2 publishes the context layer at §10.2 dimension 1 (Community risk); v1.3 introduces Derived downstream-risk score subject to methodological work surviving external review.
- Generalised per-contract EITI disclosure surface (§10.2 dimension 3). v1.1.2 ships EITI-C as a targeted disclosure-data deepening on named contexts only — Simandou (December 2025 contract disclosures) at v1.1.2 publication; Lobito Corridor (2025 EITI work) at v1.2 Phase 2 alongside the deeper Lobito Bankability + Downstream Risk panel build cycle. The 29-country EITI status-lookup remains unchanged across all other dossiers; the 30/25/25/20 country-composite weighting per §03 is unaffected. A generalised per-contract disclosure surface across the broader EITI member set remains deferred pending a user or acquirer signal that one is wanted; not dated to v1.3 specifically. Methodological precondition for any future generalised build: published per-contract field-structure standard surviving external review, plus per-publication verification discipline scaling to the full EITI member set.
- Real options analytical layer — mining investments have abandonment, expansion, and pause options; v1.4+ integration.
- AI-augmented scenario generation — v1.3 may introduce AI-augmented scenario generation as additional Derived output with same Quality Standard discipline applied.
- Probability-weighted multi-pathway NPV for disputed-tenure assets — supplementary methodology for assets carrying material πt per the Discount Rate Derivation Framework v1 §3.4; v1.3 forward-state.
- Probability-weighted structural-settlement-durability modelling — supplementary methodology for assets carrying material πd per the framework's §3.5; v1.3 forward-state.
- SAVi-mining populated valuation for Kabanga — v1.2 Phase 2. SAVi-Mining Adaptation v1 ships the methodology in-progress at v1.2.0 (cashflow-modifier structure + externality category identification published; every monetised externality value Pending). The populated SAVi-extended NPV is gated on the Kabanga ESIA externality input layer (water-resource baseline, biodiversity baseline, GHG inventory, resettlement scope, local-content commitments) which is not yet in hand. Movement from Pending to Sourced lands as a v1.2 Phase 2 audit log entry per §10.9 protocol. Manono is the committed sequential second SAVi-mining application once the methodology populated-state exists; broader generalisation across asset types is deferred indefinitely pending the methodological precondition of the populated Kabanga pilot surviving external review.
§10.9 — Audit log discipline for §10 outputs
Every §10 output surfaced on the platform creates an audit log dependency. The Audit Log Protocol per Methodology v1.1 §06 applies in full:
- Material changes to any §10 Dimension output (NPV reconciliation tolerance shift; discount rate component re-derivation; ACLED query parameter change) require a new audit log entry
- The audit log entry records what changed, why, source of authoritative correction, version marker, regression check, editorial sign-off
- Quarterly audits per the Periodic Audit Cadence include §10 outputs across the five intelligence-grade institutional dossiers
Methodology continuity disclosure. §10 publishes against the platform DCF engine state documented at
DCF Engine Validation Report 13 May 2026. The validation report records that 1 of 5 Kakula anchors reconciles within ±15% as of v1.0.54.1; the remaining 4 anchors require multi-phase capex schedule and long-term price-curve opts scheduled v1.0.55+. §10 publication does not assert reconciliation pass on assets where the engine validation has not closed; per-asset DCF outputs carry the appropriate Sourced/Derived/Pending tag at panel-render time. v1.2 publication of §10 establishes the framework; v1.3+ cycle closes remaining engine calibration gaps.