Climate VaR vs traditional VaR: comparing risk models for climate-exposed portfolios

Why It Matters

A 2025 analysis by the Network for Greening the Financial System (NGFS) found that climate-related financial losses could reach 13 percent of global GDP by 2050 under a delayed-transition scenario, yet fewer than 30 percent of institutional investors have integrated climate-adjusted risk metrics into portfolio construction (NGFS, 2025). Traditional value-at-risk (VaR) remains the dominant tool for measuring portfolio downside exposure in regulated financial institutions, but it was engineered for a world of stationary distributions and short time horizons. Climate change introduces non-stationary physical risks, abrupt policy shifts, and technological disruptions that unfold over decades rather than trading days. Climate VaR attempts to fill this gap by projecting portfolio losses under forward-looking warming and transition scenarios. The stakes are enormous: the Bank of England's 2025 climate stress test revealed that UK banks face potential credit losses of up to £225 billion under a disorderly transition pathway (Bank of England, 2025). For portfolio managers, risk officers, and regulators, understanding the structural differences between these two models is no longer optional.

Key Concepts

Traditional VaR estimates the maximum expected loss over a defined holding period at a given confidence level, typically 95 or 99 percent. It relies on historical return distributions, volatility clustering models (such as GARCH), and correlation matrices derived from market data spanning one to ten years. The three standard approaches are historical simulation, parametric (variance-covariance), and Monte Carlo simulation. Traditional VaR excels at capturing market risk, liquidity risk, and short-term tail events. Under Basel III and IV frameworks, it forms the backbone of regulatory capital calculations for trading books across global banks.

Climate VaR extends the risk framework by incorporating forward-looking scenarios that model physical climate impacts (such as extreme weather losses, sea-level rise, and chronic temperature shifts) and transition risks (such as carbon pricing, stranded assets, and technology disruption). MSCI, one of the earliest commercial providers, defines Climate VaR as the present value of costs and opportunities arising from future climate-related risks and policy changes applied to security-level cash flows (MSCI, 2025). Unlike traditional VaR, which typically uses a one-day to ten-day horizon, Climate VaR projects losses over periods of 10 to 80 years, aligned with IPCC warming pathways.

Scenario dependence is a defining feature of Climate VaR. Results vary dramatically depending on whether the modeler selects a 1.5°C orderly transition, a 2°C delayed transition, or a 3°C+ hot-house world. The NGFS provides six canonical scenarios that most commercial Climate VaR providers have adopted as reference pathways (NGFS, 2025).

Fat tails and non-stationarity. Traditional VaR assumes that past return distributions approximate future ones, an assumption that breaks down when structural shifts in energy systems, regulation, or physical geography alter asset fundamentals. Climate VaR explicitly models these non-stationary dynamics, but at the cost of far greater model uncertainty.

Stranded asset risk refers to the permanent write-down of fossil fuel reserves, carbon-intensive infrastructure, or real estate in climate-vulnerable locations. Traditional VaR captures stranded asset risk only after market prices have already declined. Climate VaR attempts to price in future stranding before it materializes in mark-to-market losses.

Head-to-Head Comparison

Physical risk modeling. Traditional VaR treats catastrophic weather events as tail risks drawn from historical frequency distributions. Climate VaR integrates forward projections from climate models such as CMIP6 to estimate how the frequency and severity of hurricanes, floods, droughts, and wildfires will shift under different warming pathways. Swiss Re (2025) estimates that insured losses from natural catastrophes reached $145 billion in 2025, a figure that backward-looking VaR models consistently underestimate because they calibrate on decades when climate extremes were less frequent.

Transition risk modeling. Carbon pricing is the primary transition risk channel in most Climate VaR implementations. MSCI's model applies projected carbon prices (ranging from $50 to $250 per tonne CO₂e by 2030 depending on scenario) to company-level Scope 1, 2, and 3 emissions to estimate future cost increases and margin compression (MSCI, 2025). Traditional VaR captures carbon price risk only for companies already subject to emissions trading systems, and only to the extent that historical carbon price volatility informs the distribution.

Validation challenge. Traditional VaR benefits from daily back-testing: institutions compare predicted losses against actual portfolio outcomes and recalibrate continuously. Climate VaR cannot be back-tested in any meaningful sense because its projections span decades. The European Central Bank (ECB, 2024) acknowledged this limitation in its supervisory guidance but argued that "the absence of back-testability does not diminish the relevance of forward-looking climate risk assessment for prudential supervision."

Real-world adoption. BlackRock integrated MSCI's Climate VaR into its Aladdin platform in 2024, enabling $10 trillion in managed assets to screen for climate-adjusted downside risk (BlackRock, 2025). Norges Bank Investment Management, which oversees Norway's $1.7 trillion sovereign wealth fund, uses an in-house Climate VaR framework to set sectoral risk budgets and inform divestment decisions (NBIM, 2025). JPMorgan Chase applies a proprietary hybrid model that layers Climate VaR outputs onto its existing historical VaR engine for its commercial real estate lending book, covering over $170 billion in exposure (JPMorgan Chase, 2025).

Key Players

Established Leaders

• MSCI — Pioneer of commercial Climate VaR, covering over 10,000 issuers with scenario-conditional loss estimates. Integrated into BlackRock's Aladdin and widely adopted by European pension funds.
• S&P Global Sustainable1 — Provides physical and transition risk scores across equities, fixed income, and real assets, powering climate stress tests for regulators and asset managers.
• Moody's Analytics (formerly Bureau van Dijk + Four Twenty Seven) — Combines physical hazard modeling with credit risk analytics, used by over 100 banks for ECB and Bank of England stress test submissions.
• Bloomberg — Offers Climate VaR and temperature alignment data through Bloomberg Terminal, covering sovereign and corporate securities.

Emerging Startups

• Intensel — Hong Kong-based startup using AI and satellite data for asset-level physical climate risk scoring across real estate and infrastructure portfolios.
• Sust Global — Provides API-first climate risk data and scenario analysis, serving asset managers and insurers with high-resolution hazard projections.
• Cervest — Offers EarthScan, a platform delivering asset-level climate intelligence for physical risk assessment across supply chains and real estate.
• Jupiter Intelligence — Delivers hyper-local climate risk analytics for physical hazard modeling, used by banks and insurers for loan and underwriting decisions.

Key Investors/Funders

• Network for Greening the Financial System (NGFS) — Central bank coalition of 134 members driving scenario standardization and supervisory expectations for climate risk modeling.
• UN Principles for Responsible Investment (PRI) — Represents over 5,000 signatories managing $120+ trillion, promoting Climate VaR adoption through reporting frameworks and signatory guidance.
• Bezos Earth Fund — Funds climate data infrastructure and open-source climate risk modeling initiatives.
• Rockefeller Foundation — Supports climate finance innovation including risk analytics for developing market portfolios.

Action Checklist

• Audit your current VaR framework. Identify which climate-related risk channels (physical hazards, carbon pricing, technology shifts, litigation) are absent from your existing models. Document blind spots in writing for your risk committee.

• Select Climate VaR scenarios aligned with your investment horizon. If your portfolio has a weighted average duration under five years, focus on disorderly transition scenarios where policy shocks arrive abruptly. For longer-duration portfolios (pensions, endowments), run all six NGFS scenarios to stress-test across warming pathways.

• Integrate asset-level geolocation data. Climate VaR accuracy depends on knowing where physical assets sit. For equities, map revenue exposure by facility location. For real estate and infrastructure, use geospatial hazard layers from providers such as Moody's or Jupiter Intelligence.

• Layer Climate VaR onto existing risk reporting. Do not replace traditional VaR. Use Climate VaR as a complementary lens that informs strategic allocation, sector tilts, and engagement priorities. Present both metrics side by side to investment committees.

• Establish a governance process for scenario selection. Avoid cherry-picking favorable scenarios. Adopt the NGFS reference scenarios as a baseline and document any deviations. Update scenario assumptions at least annually as climate science and policy evolve.

• Engage portfolio companies on emissions data quality. Climate VaR outputs are only as reliable as the underlying emissions inputs. Push for audited Scope 1, 2, and 3 disclosures aligned with ISSB S2 and the GHG Protocol. Prioritize engagement with holdings where estimated (rather than reported) emissions exceed 50 percent of the total.

• Prepare for regulatory convergence. The ECB already requires climate stress testing for supervised banks. The U.S. SEC's climate disclosure rules and the ISSB standards are driving demand for standardized Climate VaR reporting globally. Build internal capacity now to avoid compliance bottlenecks in 2027 and beyond.

FAQ

Can Climate VaR replace traditional VaR for regulatory capital calculations? Not at present. Basel III and IV frameworks mandate VaR and expected shortfall (ES) calculations based on historical market data for trading book capital requirements. Climate VaR is a supervisory and strategic tool, not a direct substitute. However, the ECB's 2024 climate stress test guidance requires banks to demonstrate that climate risks are "adequately captured in internal capital adequacy assessments," creating a de facto requirement for climate-adjusted risk metrics alongside traditional VaR (ECB, 2024). Regulatory convergence is expected to accelerate through 2028 as the Basel Committee evaluates Pillar 1 treatment of climate-related financial risks.

How reliable are Climate VaR estimates given the uncertainty in climate models? Model uncertainty is substantial. MSCI (2025) reports that Climate VaR estimates for the same portfolio can vary by a factor of three to five depending on scenario choice, discount rate, and emissions data quality. However, this uncertainty does not render the tool useless. The value lies in relative risk ranking rather than precise point estimates. A portfolio that shows consistently higher Climate VaR across multiple scenarios is genuinely more climate-exposed than one that does not, even if the absolute loss figure carries wide confidence bands. The NGFS recommends presenting results as ranges rather than single numbers.

What data gaps most undermine Climate VaR accuracy? Three data gaps dominate. First, Scope 3 emissions remain largely estimated rather than reported for most companies, and Scope 3 typically constitutes 70 to 90 percent of total value chain emissions for sectors such as finance, technology, and retail (CDP, 2025). Second, asset-level geolocation data is incomplete for complex global supply chains, limiting physical risk attribution. Third, forward-looking capex alignment data, which indicates whether companies are investing in line with transition pathways, is available for fewer than 2,000 companies globally (Transition Pathway Initiative, 2025). Improving these data inputs is the single highest-leverage action for increasing Climate VaR reliability.

Is Climate VaR relevant for fixed-income portfolios? Yes, and increasingly so. Climate VaR providers have expanded coverage from equities to corporate bonds, sovereign debt, and municipal bonds. For corporate credit, the model estimates how climate-related cost increases affect debt service capacity, probability of default, and loss given default. For sovereign debt, providers such as S&P Global model how physical climate damages and transition costs affect fiscal capacity and credit ratings. The green bond market, which reached $600 billion in annual issuance in 2025 (Climate Bonds Initiative, 2025), is driving demand for Climate VaR at the instrument level to verify that labeled bonds genuinely reduce portfolio climate risk.

How should portfolio managers communicate Climate VaR results to boards and clients? Present Climate VaR alongside traditional VaR as a complementary risk dimension, not a competing one. Use scenario narratives rather than single numbers: "Under a disorderly transition scenario, this portfolio faces an estimated 8 to 14 percent value-at-risk over a 15-year horizon, compared with 3 to 6 percent under an orderly transition." Highlight the sectors and positions that drive the largest share of Climate VaR, and connect the results to concrete actions such as engagement targets, exclusion thresholds, or hedging strategies. Visual dashboards that map physical risk by geography and transition risk by sector are more effective than dense tables.