Deep dive: Risk management & portfolio construction — what's working, what's not, and what's next

The integration of climate and sustainability considerations into portfolio risk management has moved from a niche concern of values-driven investors to a core competency demanded by regulators, fiduciaries, and asset owners managing trillions of dollars. Assets in funds incorporating some form of climate risk analysis exceeded $38 trillion globally in 2025, yet the sophistication and reliability of the underlying risk models, data inputs, and portfolio construction techniques vary enormously. Some approaches deliver genuine alpha and risk reduction. Others amount to little more than marketing overlays on conventional strategies. This deep dive evaluates what is actually working, what remains broken, and where the most important developments will occur over the next three to five years.

Why It Matters

Climate-related financial risks are no longer hypothetical. Physical risks from extreme weather events caused insured losses of $135 billion in 2024, the fourth consecutive year exceeding $100 billion. Transition risks are equally material: the European Central Bank's 2024 climate stress test found that a disorderly transition scenario could generate credit losses of 8-10% across European bank portfolios, concentrated in carbon-intensive sectors. Asset stranding analysis by Carbon Tracker estimates that $1.4 trillion in fossil fuel assets could become uneconomic under a 1.5 degree pathway by 2030.

Regulatory pressure compounds the financial imperative. The EU's Sustainable Finance Disclosure Regulation (SFDR) requires fund managers to classify products and disclose sustainability risks. The UK's Transition Plan Taskforce framework mandates that large companies and financial institutions publish credible transition plans. The SEC's climate disclosure rules, while subject to legal challenges, establish expectations for how public companies quantify and report climate-related financial risks. Japan's Financial Services Agency requires TCFD-aligned reporting from listed companies, and Singapore's MAS has integrated climate risk into supervisory expectations for banks and insurers.

For sustainability leads at asset managers, pension funds, and corporate treasury functions, the operational question is no longer whether to integrate climate risk but how to do it in ways that improve portfolio outcomes rather than simply generating compliance documentation. The gap between leaders and laggards is widening, and the consequences of getting this wrong, whether through naive exclusion strategies that sacrifice returns, or through inadequate risk assessment that leaves portfolios exposed, are increasingly quantifiable.

Key Concepts

Climate Value-at-Risk (CVaR) extends traditional Value-at-Risk frameworks to estimate potential portfolio losses from climate-related physical and transition risks over specified time horizons. Leading providers including MSCI, Moody's, and S&P Global offer CVaR models that assess company-level exposure to carbon pricing, technology shifts, market sentiment changes, extreme weather, and chronic physical changes. The output typically estimates the percentage of enterprise value at risk under different warming scenarios (1.5, 2.0, and 3.0+ degree pathways). Current models exhibit significant divergence: a 2024 analysis by the Bank for International Settlements found that CVaR estimates for the same portfolio varied by a factor of three across providers, reflecting differences in scenario assumptions, transition speed modeling, and physical risk parameterization.

Scenario Analysis evaluates portfolio performance under multiple plausible future states rather than optimizing for a single expected outcome. The Network for Greening the Financial System (NGFS) provides standardized climate scenarios used by over 130 central banks and supervisors globally. These scenarios model interactions between climate policy, technology, energy markets, and macroeconomic variables across orderly transition, disorderly transition, and hot house world pathways. Effective portfolio construction uses scenario analysis to identify strategies that perform reasonably well across multiple futures rather than concentrating bets on a single pathway.

Transition Pathway Assessment evaluates whether individual companies have credible plans to decarbonize their operations and value chains in line with climate targets. The Transition Pathway Initiative (TPI), used by investors managing over $50 trillion in assets, scores companies on management quality and carbon performance relative to sector benchmarks. Paris Agreement Capital Transition Assessment (PACTA) provides portfolio-level analysis of alignment with temperature pathways. These tools inform portfolio construction by distinguishing between companies genuinely transitioning and those engaged in incremental improvements insufficient to avoid stranded asset risk.

Physical Risk Scoring quantifies asset-level exposure to acute hazards (floods, hurricanes, wildfires, heatwaves) and chronic changes (sea level rise, water stress, agricultural yield shifts). Providers including Four Twenty Seven (now part of Moody's), Jupiter Intelligence, and XDI map physical risks to specific coordinates, enabling granular assessment of real estate, infrastructure, and supply chain vulnerabilities. Physical risk scoring has become particularly important for real asset portfolios, insurance-linked securities, and municipal bond analysis.

What's Working

Factor Integration in Equity Portfolios

The most evidence-backed approach integrates climate risk as an additional factor within systematic equity strategies rather than treating it as a standalone overlay. Research from MSCI and Robeco demonstrates that portfolios optimized to reduce carbon intensity by 30-50% while maintaining sector, size, and style factor exposures have delivered tracking errors below 1% relative to parent benchmarks over 5-year periods, meaning investors can materially reduce carbon exposure with minimal performance impact.

The mechanism works because within most sectors, lower-carbon companies tend to have operational characteristics (energy efficiency, newer capital stock, stronger management) that correlate with quality and profitability factors. AXA Investment Managers' climate-aware equity range, managing over EUR 40 billion, has demonstrated this approach at scale, achieving 50%+ carbon intensity reduction versus benchmarks with competitive risk-adjusted returns. The key design principle is optimization rather than exclusion: instead of removing high-carbon companies entirely, the portfolio tilts toward lower-carbon alternatives within each sector and region, preserving diversification.

Sovereign Bond Climate Risk Assessment

Climate risk integration in sovereign debt, historically lagging equity approaches, has matured significantly. The NGFS Sovereign Climate Risk Dashboard, released in 2024, provides standardized physical and transition risk scores for 180+ countries. Research from the University of Cambridge demonstrates that sovereign credit spreads already partially price climate vulnerability, with physically exposed nations (small island developing states, low-lying coastal economies) paying 50-200 basis points more than comparable-risk peers after controlling for conventional macro variables.

Asset managers including PIMCO, Amundi, and Nordea have operationalized sovereign climate risk scoring into fixed income portfolio construction. The practical implementation involves adjusting duration and country allocation based on climate-adjusted risk-return profiles. For emerging market debt portfolios, where climate physical risks are most acute, this has improved downside protection during extreme weather events while maintaining yield targets.

Private Market Due Diligence

Climate risk integration in private equity and infrastructure due diligence has advanced substantially because the illiquid, long-duration nature of these investments makes climate risk especially material. Infrastructure investors now routinely conduct asset-level physical risk assessments for energy, transport, and water assets with economic lives of 25-50 years. Brookfield Asset Management's transition investment program, with $28 billion committed, applies proprietary climate risk scoring to every potential acquisition, adjusting valuation models for carbon pricing trajectories, stranded asset risk, and physical hazard exposure.

The value is demonstrable: CPPIB (now CPP Investments) reported that integrating climate risk into infrastructure investment decisions avoided three investments in 2023-2024 that subsequently experienced material value impairment from regulatory or physical risk events. Private credit managers are similarly integrating transition risk into underwriting standards, with Ares Management and Blackstone Credit requiring climate risk assessments for energy-exposed borrowers.

What's Not Working

Overreliance on Carbon Metrics as Risk Proxies

Portfolio carbon footprinting remains the most widely adopted climate risk tool, yet carbon intensity is a poor proxy for financial risk. A portfolio's weighted average carbon intensity (WACI) measures current emissions but reveals little about forward-looking transition risk, physical risk exposure, or management quality. Divesting from a well-managed utility investing heavily in renewable generation (high current emissions, declining trajectory) while retaining exposure to a consumer goods company with significant deforestation supply chain risk (low direct emissions, growing regulatory exposure) represents a common failure mode.

The overemphasis on Scope 1 and 2 emissions data further distorts portfolio construction. For most sectors, including technology, financial services, apparel, and food, Scope 3 value chain emissions represent 80-95% of total emissions. Scope 3 data quality remains poor, with estimates varying by 40-60% across providers for the same company. Portfolio managers constructing strategies based primarily on Scope 1 and 2 data are optimizing for measurability rather than materiality.

Model Divergence and False Precision

The divergence across climate risk model providers undermines confidence and actionability. A 2025 study by the European Securities and Markets Authority (ESMA) found that ESG ratings from major providers exhibited correlations of only 0.45-0.65, compared to credit ratings, which correlate at 0.95+. Climate-specific metrics show even greater variance. Temperature alignment scores for the same company range from 1.8 to 3.5 degrees depending on the provider, rendering the metric nearly meaningless for precise portfolio construction.

The root causes include: divergent methodological choices (top-down vs. bottom-up emissions accounting), inconsistent boundary definitions (which Scope 3 categories to include), different scenario assumptions (carbon price trajectories, technology adoption curves), and varying approaches to missing data (estimation vs. exclusion). The resulting "model uncertainty" often exceeds the "climate signal" being measured, creating a false sense of analytical rigor.

Short-Horizon Bias in Risk Models

Most institutional risk management frameworks operate on 1-3 year horizons aligned with performance evaluation cycles and reporting periods. Climate risks, particularly physical risks from gradual warming and transition risks from long-term policy shifts, materialize over 5-30 year horizons. This mismatch means that conventional portfolio risk tools systematically underweight climate risks because they fall outside the measurement window.

The consequence is a structural bias toward inaction. Quarterly risk reports show minimal climate-related variance, reinforcing the perception that climate risks are not yet material. By the time risks manifest within conventional measurement horizons, repricing is abrupt and portfolio damage is concentrated. The failure of several US regional banks with concentrated commercial real estate exposure in climate-vulnerable geographies during 2024-2025, which was predictable from a 10-year physical risk assessment but invisible in quarterly risk metrics, illustrates this dynamic.

What's Next

Integrated Physical-Transition Risk Modeling

The next generation of climate risk models will integrate physical and transition risks into unified frameworks that capture their interactions. Currently, most providers model physical and transition risks independently, missing critical feedback loops. Severe physical events accelerate policy responses (transition risk), while aggressive transition policies reduce long-term physical risks but create near-term economic disruptions. Unified models from emerging providers including Planetrics (now part of McKinsey) and Ortec Finance are beginning to capture these dynamics, enabling more nuanced portfolio construction that accounts for correlated scenarios.

Nature and Biodiversity Risk Integration

The Taskforce on Nature-related Financial Disclosures (TNFD) framework, with over 400 institutional adopters by early 2026, is extending portfolio risk analysis beyond climate to encompass biodiversity loss, ecosystem degradation, water scarcity, and land-use change. Early analysis by the Dutch central bank (DNB) found that 36% of Dutch financial institution assets have material biodiversity-related dependencies. Portfolio construction will increasingly need to assess nature-related risks using frameworks that parallel climate risk analysis but require different data sources (spatial biodiversity data, ecosystem service mapping) and models.

Regulatory Convergence on Transition Plans

The proliferation of transition plan requirements (UK TPT, EU CSRD, ISSB) is creating a standardized information base that will improve portfolio construction. When companies publish detailed, comparable transition plans with interim targets, capital expenditure plans, and governance structures, portfolio managers can more accurately assess transition credibility and price transition risk. This shift from backward-looking emissions data to forward-looking strategic assessment will fundamentally improve the quality of climate-informed portfolio construction.

AI-Enhanced Scenario Generation

Machine learning is beginning to supplement traditional integrated assessment models for climate scenario generation. AI systems can process vastly more data inputs, including real-time policy signals, patent filings, corporate capital expenditure announcements, and commodity price movements, to generate higher-frequency, more granular scenarios. These tools will enable dynamic portfolio adjustment based on evolving transition signals rather than annual rebalancing based on static scenarios.

Action Checklist

• Audit current climate risk tools and data providers, mapping known gaps between measured metrics and material financial risks
• Extend risk measurement horizons beyond standard 1-3 year windows to include 5-10 year climate scenario analysis
• Shift from carbon-intensity-only metrics to multi-dimensional risk assessment incorporating transition credibility, physical exposure, and Scope 3 emissions
• Benchmark climate risk model outputs across at least two providers to identify divergence and calibrate confidence levels
• Integrate physical risk scoring at the asset level for real estate, infrastructure, and geographically concentrated exposures
• Develop transition pathway assessment capabilities for high-emitting sector holdings using TPI and PACTA frameworks
• Prepare for TNFD integration by inventorying portfolio dependencies on nature-related ecosystem services
• Engage with portfolio companies on transition plan quality, using emerging regulatory frameworks as benchmarks for credibility
• Train investment teams on climate scenario analysis interpretation and portfolio construction implications

Sources

• Bank for International Settlements. (2024). Climate Risk Modelling: A Cross-Provider Comparison and Reliability Assessment. Basel: BIS.
• European Central Bank. (2024). Economy-Wide Climate Stress Test: Results and Methodology. Frankfurt: ECB Publications.
• Carbon Tracker Initiative. (2025). Stranded Assets and the Fossil Fuel Industry: 2025 Update. London: Carbon Tracker.
• MSCI. (2025). Climate Risk in Portfolio Construction: Factor-Based Approaches and Performance Evidence. New York: MSCI Research.
• Network for Greening the Financial System. (2024). NGFS Climate Scenarios for Central Banks and Supervisors: Technical Documentation. Paris: NGFS.
• Transition Pathway Initiative. (2025). State of Transition Report 2025. London: Grantham Research Institute, London School of Economics.
• European Securities and Markets Authority. (2025). ESG Ratings: Market Structure and Provider Divergence. Paris: ESMA.
• Taskforce on Nature-related Financial Disclosures. (2025). Recommendations and Guidance for Financial Institutions. Geneva: TNFD Secretariat.