TNFD for Insurers: Underwriting, Investments and Nature-Related Risk

Educational practitioner material. Illustrative scenarios and wording require adaptation to the organisation’s facts, materiality approach, sector, locations and reporting context.

Insurance businesses can be connected to nature in three materially different ways. Underwriting creates exposure to insured assets, activities, liabilities and business interruption. Investments connect the balance sheet to issuers, projects, sovereigns and real assets. Own operations create a third, usually smaller, boundary covering offices, facilities, procurement and the claims supply chain. Combining all three too early can hide who owns the risk, which data are available and what management response is possible.

The practical objective is not to calculate one universal nature score. It is to identify credible pathways from an insured or invested activity and location to an ecosystem dependency or impact, then to physical, transition, liability or systemic consequences for claims, premiums, reserves, asset values, operating costs, product demand and access to reinsurance or capital. The same exposure can create different pathways on the asset and liability sides of the insurer.

Figure 1. A three-lens operating model for insurers. Underwriting, investments and own operations are assessed separately and then reconciled through shared governance, scenarios, controls and evidence.

A line-of-business code is useful for screening, but the risk pathway usually depends on the underlying activity and place. Property insurance for a food processor, liability insurance for an agrochemical manufacturer and marine cover for a port may all sit in familiar product classes while depending on very different ecosystems and impact drivers. The underwriting data model therefore needs a bridge from policy and insured party to activity, asset, location, revenue dependency and relevant value chain.

Catastrophe models are inputs, not the whole answer

Catastrophe models can provide strong evidence for acute hazards and insured loss. They may also incorporate land cover, drainage or coastal conditions. However, an insurer should test what the model does not capture: chronic degradation of water, soil or ecosystem condition; ecological thresholds; policy and liability changes; client adaptation; availability of repair materials; claims inflation; and correlations across lines. Model version, spatial resolution, event set, vulnerability assumptions and treatment of ecosystem change should remain visible in the evidence pack.

Liability and claims pathways

Nature-related liability risk can arise from alleged or actual environmental harm, failure to meet duties, inadequate disclosure, professional advice, directors responsibilities, remediation or compensation. The underwriting team should not treat every environmental controversy as a claim forecast. It should identify the insured duty, plausible claimant, trigger, coverage interaction, defence and remediation pathway, time horizon and uncertainty. Claims and legal teams are important sources of emerging-risk evidence because loss development may extend beyond the annual underwriting cycle.

For investments, the insurer can begin with sector and geography screening, but material conclusions need more specific information. The portfolio register should retain asset class, issuer or project, exposure amount, industry classification, operating regions, known sites or commodities, data source, confidence and date. Priority holdings can then be assessed using the relevant real-economy sector guidance, issuer engagement and scenario analysis.

Insurance decisions operate over several time horizons: a one-year policy period, a multi-year treaty or infrastructure project, a long-tail liability claim, and a long-duration investment. Scenario analysis should therefore test more than one date and should link nature and climate uncertainties. A practical first cycle can use qualitative narratives to test portfolio concentration, claims severity, pricing adequacy, reinsurance availability, asset values and response options before attempting granular loss quantification.

1. Lock the entity and portfolio boundary. List underwriting entities, books, investment pools, own operations, periods, exclusions and accountable owners.

2. Screen activities and locations. Use sector, commodity, geography and hazard information to identify potentially high-dependency and high-impact populations.

3. Prioritise portfolios for deeper work. Combine exposure size, nature interface, sensitive locations, claims relevance, strategic importance and data confidence.

4. Build DIRO pathways. For each priority population, connect dependencies and impacts to triggers, business consequences and financial effects.

5. Reconcile catastrophe and nature data. Document model coverage and supplement missing chronic, transition, liability and ecosystem-service information.

6. Assess controls and responses. Test underwriting guidelines, policy terms, risk engineering, client engagement, investment stewardship and internal limits.

7. Run proportional scenarios. Use at least one physical-degradation and one transition-response pathway over relevant policy, claims and investment horizons.

8. Select metrics and targets. Choose portfolio coverage, risk, opportunity, engagement and own-operations metrics that represent material issues without false precision.

9. Integrate governance and finance. Connect conclusions to risk appetite, product governance, reserving assumptions, capital allocation, investment committees and board reporting.

10. Draft with limitations. Explain scope, methods, estimates, confidence, changes, confidential information and the expansion plan.

Illustrative scenario. A composite insurer covers food processors, farms and logistics assets in one basin and holds bonds issued by several of the same groups. Chronic groundwater decline raises production cost and business interruption probability. A pollution incident then leads to clean-up, liability claims and tighter permit conditions. The insurer initially sees separate property, environmental liability and credit exposures. The TNFD assessment reveals a shared ecosystem dependency and correlated physical, transition and investment pathways.

The insurer does not immediately assign a precise total loss. It first maps insured locations and investment exposures, reviews model coverage, identifies the policy terms and claims histories that matter, tests a severe but plausible basin scenario, and records where data are proxy-based. Responses include accumulation limits, water-risk engineering, revised referral rules, engagement with investees and clients, and monitoring of permit and ecosystem indicators. The disclosure explains that quantitative potential-loss estimates remain under development.