Case study: Climate risk stress testing & scenario regulation — a leading company's implementation and lessons learned

When HSBC Holdings published its first climate stress test results in 2023, the bank disclosed that a disorderly net-zero transition scenario could generate up to $4.3 billion in additional expected credit losses across its global lending portfolio by 2050. That figure, which exceeded many analysts' expectations, marked a turning point for climate risk stress testing in European banking. The Bank of England's 2022 Climate Biennial Exploratory Scenario (CBES) had already revealed that the UK's largest banks and insurers could face combined losses of up to 10 to 15% of annual profits under adverse climate scenarios (Bank of England, 2022). For financial institutions navigating an increasingly prescriptive regulatory environment, these results underscore a hard truth: climate risk stress testing is no longer a voluntary exercise but a core prudential requirement with direct balance-sheet implications.

Why It Matters

Climate risk stress testing sits at the intersection of financial regulation, climate science, and corporate strategy. Regulators in the UK, EU, US, and Asia-Pacific have moved from publishing general supervisory expectations to mandating specific stress testing exercises with defined scenarios, time horizons, and reporting requirements. The European Central Bank's (ECB) 2022 supervisory climate stress test covered 104 significant institutions and found that 60% lacked adequate climate risk data infrastructure, while only 36% had integrated climate scenarios into their internal capital adequacy assessment processes (ECB, 2022).

The financial stakes are substantial. The Network for Greening the Financial System (NGFS) estimates that delayed climate action could reduce global GDP by up to 13% by 2050 under a "hot house world" scenario, translating into trillions of dollars in potential asset impairments across the financial system (NGFS, 2023). For individual institutions, failure to demonstrate robust stress testing capabilities risks supervisory action, capital add-ons, and reputational damage in a market where investors increasingly scrutinize climate risk management quality.

In the UK specifically, the Prudential Regulation Authority (PRA) has set clear expectations through Supervisory Statement SS3/19, requiring banks and insurers to embed climate risk into governance, risk management, scenario analysis, and disclosure. Institutions that treat stress testing as a compliance checkbox rather than a strategic capability will find themselves at a competitive disadvantage as the regulatory bar continues to rise.

Key Concepts

Climate risk stress testing evaluates the resilience of a financial institution's balance sheet under hypothetical but plausible climate scenarios. Physical risk scenarios model the financial impact of acute events (floods, storms, wildfires) and chronic changes (sea-level rise, temperature increases) on asset values, creditworthiness, and insurance liabilities. Transition risk scenarios model the impact of policy changes (carbon pricing, fossil fuel phase-outs), technology shifts (renewable energy cost declines, EV adoption), and market sentiment changes on sectors exposed to decarbonization.

The NGFS provides a set of reference scenarios that most regulators have adopted as a starting framework. These range from "orderly transition" (early, gradual policy action) through "disorderly transition" (late, abrupt policy action) to "hot house world" (insufficient action, severe physical impacts). Institutions are expected to translate these macro scenarios into sector-specific and geography-specific impacts on their individual portfolios.

Scenario analysis differs from traditional financial stress testing in several important ways: time horizons extend to 2050 or 2080, far beyond typical 3 to 5 year planning cycles; historical data provides limited guidance because the scenarios describe unprecedented structural shifts; and the transmission channels from climate drivers to financial outcomes involve complex, non-linear physical and economic dynamics that require specialized modeling capabilities.

What's Working

HSBC's Enterprise-Wide Climate Stress Testing Program

HSBC's implementation of climate stress testing offers one of the most comprehensive examples of a global bank building this capability at scale. Beginning in 2021, HSBC assembled a dedicated climate risk analytics team of approximately 40 professionals spanning climate science, credit risk modeling, data engineering, and regulatory reporting. The team's mandate was to develop an end-to-end capability that could assess climate-related credit, market, and operational risks across HSBC's $2.4 trillion balance sheet.

The bank adopted a phased approach. Phase one (2021 to 2022) focused on the Bank of England's CBES exercise, which required analysis of three scenarios across a 30-year time horizon for the UK lending book. Phase two (2022 to 2023) extended coverage to the full global portfolio, incorporating the ECB's supervisory stress test requirements and Hong Kong Monetary Authority expectations. Phase three (2023 to 2025) embedded scenario analysis into business-as-usual credit risk processes, including sector-level transition risk assessments for high-emitting industries and physical risk scoring for commercial real estate exposures.

Key results from HSBC's program include identification of $85 billion in lending exposures to sectors classified as "transition-sensitive" (oil and gas, thermal coal, heavy industry, aviation, and shipping), development of property-level physical risk scores for 1.2 million commercial and residential properties using geospatial climate hazard data, and quantification of expected credit loss impacts under multiple NGFS scenarios that informed capital planning discussions with the PRA.

Aviva's Insurance-Side Stress Testing

Aviva, one of the UK's largest insurers with approximately $400 billion in assets under management, implemented climate stress testing across both its insurance liabilities and investment portfolio. The insurer's approach illustrates how stress testing can drive tangible business decisions rather than remain a purely analytical exercise.

Aviva's physical risk stress testing analyzed the impact of increasing flood, windstorm, and subsidence frequency on its UK general insurance book. Using scenario-adjusted catastrophe models from RMS (now Moody's RMS) and AIR Worldwide, the insurer quantified potential increases in claims costs under 2 degrees Celsius and 4 degrees Celsius warming pathways. The analysis revealed that UK flood losses could increase by 25 to 40% under a 2 degrees Celsius scenario and by 60 to 90% under a 4 degrees Celsius scenario by 2050, driven primarily by increased precipitation intensity and sea-level rise affecting coastal and fluvial flood zones (Aviva, 2024).

On the investment side, Aviva stress-tested its $60 billion credit portfolio against transition scenarios, identifying that approximately 8% of holdings were in sectors facing material stranded-asset risk under orderly transition pathways, rising to 15% under disorderly scenarios. These results directly informed portfolio rebalancing decisions, with Aviva accelerating its exit from thermal coal exposures and increasing allocations to green bonds and climate solution equities.

Standard Chartered's Emerging Market Focus

Standard Chartered Bank's climate stress testing program is notable for its focus on emerging markets, where the bank generates approximately 90% of its revenue. The bank developed bespoke physical risk models for markets including India, Bangladesh, Indonesia, and Nigeria, where exposure to extreme heat, flooding, and tropical cyclones intersects with high credit concentrations in agriculture, infrastructure, and real estate.

Standard Chartered's analysis found that physical climate risks in its Asian markets could increase probability of default by 15 to 30% for agriculture-sector borrowers under high-warming scenarios, with flood risk in Bangladesh representing the single largest physical risk concentration in the bank's global portfolio. The bank used these findings to develop climate-adjusted credit risk policies, including enhanced due diligence requirements for lending in high-risk flood zones and preferential financing terms for borrowers implementing physical adaptation measures (Standard Chartered, 2024).

What's Not Working

Data gaps remain the most significant obstacle. Most financial institutions lack granular, asset-level emissions data for their lending and investment portfolios. The ECB's 2022 assessment found that banks relied on estimated or proxy data for more than 70% of their financed emissions calculations, introducing substantial uncertainty into stress test results (ECB, 2022). Property-level physical risk data is similarly incomplete, with many institutions unable to geocode individual collateral assets to the precision required for flood, wildfire, or storm surge modeling.

Scenario translation presents ongoing challenges. The NGFS reference scenarios provide macro-level pathways, but translating these into sector-specific, geography-specific, and counterparty-specific financial impacts requires assumptions and modeling choices that can significantly influence results. Two institutions analyzing the same portfolio under the same NGFS scenario can produce markedly different loss estimates depending on their chosen transmission mechanisms, sector classification approaches, and time-step granularities. This lack of comparability undermines the supervisory value of cross-institutional stress testing.

Short-termism in governance structures limits the strategic impact of stress test findings. Climate scenarios with 2050 or 2080 horizons generate outputs that extend well beyond the tenure of current management teams and board members. Multiple institutions have reported internally that stress test results, while acknowledged at board level, have had limited influence on near-term capital allocation, lending strategy, or risk appetite decisions because the losses are projected to materialize over decades rather than quarters.

Model validation for climate risk models remains underdeveloped. Traditional model validation frameworks rely on backtesting against historical outcomes, but climate stress test models describe future scenarios with no historical precedent. The PRA has acknowledged this challenge, noting that existing model risk management frameworks "may need to be adapted" for climate models, but has not yet provided definitive guidance on acceptable validation approaches (PRA, 2023).

Key Players

Established Institutions

• HSBC Holdings: deployed enterprise-wide climate stress testing across a $2.4 trillion balance sheet, setting an industry benchmark for scope and granularity
• Aviva: integrated climate scenario analysis into both insurance underwriting and investment portfolio management decisions
• Standard Chartered: developed bespoke physical risk models for emerging market exposures across Asia and Africa
• Bank of England: designed and executed the CBES exercise, the first major central bank climate stress test
• European Central Bank: conducted supervisory climate stress testing across 104 significant eurozone institutions

Technology and Data Providers

• Moody's RMS: provides scenario-adjusted catastrophe models used by insurers for physical risk stress testing
• MSCI: offers climate value-at-risk analytics and transition risk scoring for investment portfolios
• S&P Global Sustainable1: delivers physical and transition risk datasets at the asset and counterparty level
• Ortec Finance: supplies climate scenario modeling tools adopted by pension funds and insurers

Investors and Conveners

• Network for Greening the Financial System (NGFS): develops the reference scenarios used by regulators globally
• Climate Financial Risk Forum (CFRF): the UK's cross-industry body co-chaired by the PRA and FCA that develops practical climate risk guidance
• UN Environment Programme Finance Initiative (UNEP FI): coordinates pilot programs for climate stress testing methodologies across banks and insurers

Action Checklist

• Establish a dedicated climate risk analytics function with expertise spanning climate science, credit risk modeling, and data engineering
• Map portfolio exposures to NGFS reference scenarios at sector and geography level, starting with the highest-emitting sectors and most physically exposed geographies
• Invest in asset-level data infrastructure, including geocoding of collateral assets for physical risk assessment and counterparty-level emissions data collection
• Develop sector-specific transition risk scorecards that can be embedded into credit origination and portfolio monitoring processes
• Engage with regulators proactively to understand evolving expectations and shape emerging supervisory approaches
• Integrate stress test outputs into capital planning, risk appetite, and strategic planning discussions rather than treating them as standalone compliance exercises
• Establish a model validation framework adapted for climate risk models, incorporating expert judgment, sensitivity analysis, and cross-model benchmarking
• Publish stress test results and methodology transparently to build market confidence and contribute to industry-wide capability development

FAQ

Q: How long does it typically take to build a climate stress testing capability from scratch? A: Most institutions report 18 to 36 months from initial scoping to production-quality stress test outputs. The first 6 to 12 months are typically consumed by data infrastructure development, including emissions data collection, asset geocoding, and scenario database construction. The modeling and analytics phase requires an additional 6 to 12 months, followed by 3 to 6 months for governance integration, documentation, and regulatory engagement. HSBC's phased approach over approximately 36 months is representative of a large, complex institution building comprehensive capability.

Q: What level of investment is required for a climate stress testing program? A: Costs vary significantly by institution size and complexity. For a mid-sized European bank with $100 to $500 billion in assets, initial build costs typically range from $3 million to $10 million, covering staff hiring, data procurement, technology infrastructure, and external consulting support. Ongoing annual operating costs of $2 million to $5 million are typical for maintaining data feeds, model updates, and dedicated analytics teams. Technology and data vendor costs (Moody's RMS, MSCI, S&P Global) typically represent 20 to 30% of total program expenditure.

Q: How should institutions handle the uncertainty inherent in long-term climate scenarios? A: Leading practitioners use multiple scenarios rather than a single central case, focusing on the range of potential outcomes rather than point estimates. Sensitivity analysis around key assumptions (carbon price trajectories, technology adoption rates, physical hazard intensities) helps identify which variables drive the largest differences in financial outcomes. Communicating results as ranges with explicit uncertainty bands, rather than precise figures, builds credibility with both internal stakeholders and regulators while honestly reflecting the state of scientific and economic knowledge.

Q: Are climate stress test results actually influencing business decisions today? A: Evidence is mixed. Stress testing has demonstrably influenced portfolio-level decisions at institutions such as Aviva (accelerating coal exit), and sector-level risk policies at Standard Chartered (enhanced due diligence in flood-exposed geographies). However, the Bank of England's CBES feedback noted that most participating institutions had not yet integrated scenario analysis into day-to-day credit decisions or capital allocation. The institutions making the fastest progress are those where climate stress testing is embedded within existing enterprise risk management frameworks rather than siloed in sustainability or strategy teams.

Sources

• Bank of England. (2022). Results of the 2021 Climate Biennial Exploratory Scenario. London: Bank of England.
• European Central Bank. (2022). 2022 Climate Risk Stress Test: Results and Key Findings. Frankfurt: ECB Banking Supervision.
• Network for Greening the Financial System. (2023). NGFS Scenarios for Central Banks and Supervisors: Phase IV. Paris: NGFS Secretariat.
• Aviva. (2024). Climate-Related Financial Disclosure Report 2023. London: Aviva plc.
• Standard Chartered. (2024). Climate Risk Management: Approach, Methodology, and Results. London: Standard Chartered PLC.
• Prudential Regulation Authority. (2023). Thematic Feedback: Managing Climate-Related Financial Risks (SS3/19 Implementation). London: Bank of England PRA.
• HSBC Holdings. (2024). Climate Risk: Stress Testing Methodology and Results. London: HSBC Holdings plc.
• MSCI. (2025). Climate Value-at-Risk: Methodology and Application for Financial Institutions. New York: MSCI Inc.