Risk Modelling Portfolio

Twelve modelling modules across climate, banking, insurance, ALM, IRRBB, valuation, capital and validation — each documented by objective, methodology, data, calibration, outputs, users and regulatory relevance.

Climate Physical Risk

Objective: Quantify hazard-driven financial impact from flood, wind, heat, wildfire, drought and coastal perils.
Methods: Hazard maps, event sets, exposure geocoding, vulnerability curves, event and annual loss aggregation.
Data: Public hazard datasets, elevation, land use, exposure inventories, EPC and property attributes.
Calibration: Fitted damage functions with documented parameters; benchmarked against catastrophe vendor curves.
Outputs: Event and annual losses, tail metrics, portfolio heatmaps.
Users: Underwriters, credit officers, CRO, risk committees.
Regulatory: Solvency II NatCat, PRA climate expectations, ORSA.

Physical risk articles

Climate Transition Risk

Objective: Translate transition pathways into sector-level financial impact on credit, valuation and reserves.
Methods: Carbon price shocks, sector pathway mapping, cash-flow re-modelling, credit migration.
Data: NGFS scenarios, emissions intensity, sector production, energy mix.
Calibration: Scenario-consistent shocks with documented linkages to accounting metrics.
Outputs: Portfolio PD/LGD shifts, valuation impacts, sector heatmaps.
Users: Credit risk, portfolio management, investment risk.
Regulatory: IFRS 9 macro overlays, ICAAP, ORSA transition scenarios.

Transition risk articles

IRRBB and ALM

Objective: Measure interest-rate and liquidity risk in the banking book across curves, tenors and behavioural assumptions.
Methods: EVE and NII sensitivities, parallel and non-parallel shocks, behavioural models for non-maturity deposits and prepayment.
Data: Cashflow ladders, deposit history, prepayment history, funding curves.
Calibration: Empirical behavioural parameters with sensitivity ranges and expert overrides where data is thin.
Outputs: EVE, NII, hedge effectiveness, liquidity stress metrics.
Users: Treasury, ALCO, risk committees.
Regulatory: EBA IRRBB, ILAAP, LCR, NSFR.

Credit Risk and IFRS 9

Objective: Expected credit loss modelling with forward-looking macro overlays including climate.
Methods: PD term structures, LGD, EAD, staging, macroeconomic satellites, climate overlays.
Data: Internal default and recovery data, macro time series, sector datasets.
Calibration: Statistical fitting with challenger models and out-of-sample validation.
Outputs: Stage 1/2/3 provisions, sensitivity, scenario probability weighting.
Users: Finance, credit risk, audit.
Regulatory: IFRS 9, ICAAP stress testing, EBA guidelines.

Insurance Pricing

Objective: Technical pricing across property, motor, health, specialty and life lines.
Methods: GLMs, GAMs, gradient boosting, hierarchical models; segmentation, elasticity, uplift.
Data: Claims, policy, exposure, external hazard and behavioural data.
Calibration: Backtesting on holdout, monitoring loss ratio and lift.
Outputs: Technical rates, risk selection, portfolio mix.
Users: Actuaries, underwriters, pricing committees.
Regulatory: Consumer duty, IDD, local pricing rules.

Reserving

Objective: Best estimate and risk margin under IFRS 17 and Solvency II.
Methods: Chain ladder, Bornhuetter-Ferguson, GLMs on triangles, stochastic reserving, cashflow projections.
Data: Development triangles, exposure, external claims environment.
Calibration: Backtested against actual runoff; documented outlier treatment.
Outputs: Best estimate, risk margin, sensitivity, CSM.
Users: Actuarial reserving, finance, audit.
Regulatory: IFRS 17, Solvency II TP, local statutory frameworks.

Solvency II Economic Capital

Objective: SCR and internal-model style economic capital.
Methods: Modular capital, aggregation with correlation matrices or copulas, USP and PIM approaches.
Data: Balance sheet, asset composition, underwriting distributions.
Calibration: Distributional fitting with tail focus; documented dependence structure.
Outputs: SCR, own funds, coverage ratio, sensitivities.
Users: Risk function, capital committee.
Regulatory: Solvency II Pillar 1/2, ORSA.

Basel III Capital Modelling

Objective: RWA and capital modelling across credit, market and operational risk.
Methods: IRB models, standardised, FRTB, ORC and AMA-style approaches.
Data: Portfolio-level exposure and risk parameters.
Calibration: Parameter estimation with regulator-aligned floors and add-ons.
Outputs: RWA, capital ratios, stress capital.
Users: Regulatory reporting, treasury, CRO.
Regulatory: Basel III/IV, EBA, PRA.

Liquidity Risk

Objective: Institution-level liquidity risk under normal and stressed conditions.
Methods: LCR, NSFR, behavioural liquidity, cash-flow-at-risk, funding concentration.
Data: Cashflow ladders, deposit behaviour, funding sources, HQLA.
Calibration: Historical run-rate calibration, tail scenarios.
Outputs: Coverage ratios, liquidity horizon, contingency needs.
Users: Treasury, ALCO.
Regulatory: ILAAP, LCR, NSFR.

Cross-Asset Valuation

Objective: Coherent scenario re-pricing of fixed income, equity, real estate and infrastructure.
Methods: Discount curves, spread models, factor models for equity/property, no-arbitrage constraints.
Data: Market data, macro pathways, sector performance.
Calibration: Market-consistent calibration with challenger comparisons.
Outputs: Scenario NAVs, MTM stresses, hedge effectiveness.
Users: Investment risk, ALM, treasury.
Regulatory: Solvency II SII asset stress, ICAAP market shocks.

Model Validation

Objective: Independent second-line validation across risk models.
Methods: Conceptual soundness, data quality, implementation review, backtesting, benchmarking, sensitivity.
Data: Model documentation, source code, input data, outputs.
Calibration: N/A — validation, not development.
Outputs: Findings, severity ratings, remediation plans, sign-off.
Users: Model risk function, audit, regulators.
Regulatory: SR 11-7, EBA model management, PRA SS 1/23.

Model validation notes

AI Model Governance

Objective: Governance of ML and generative components across the modelling estate.
Methods: Model inventory, explainability, monitoring, human-in-the-loop, red-team review.
Data: Training/validation splits, drift indicators, feedback loops.
Calibration: Ongoing retraining schedules with challenger champions.
Outputs: Approved use, monitoring dashboards, incident reports.
Users: Model risk, compliance, technology, business owners.
Regulatory: EU AI Act, PRA SS 1/23, SR 11-7.

AI for climate risk

Discuss a modelling engagement

Advisory, build or independent validation across any of these modules.

Contact Jonas