Why do ESG ratings from different agencies diverge so much, and what does this mean for risk management?

ESG rating divergence is one of the most significant challenges in sustainable finance. Academic research shows correlations of 0.38-0.71 between major ESG raters, compared to 0.99+ for credit ratings. This divergence stems from fundamental differences in scope, measurement, and weighting.\n\nThree Sources of Divergence (Berg, Kolbel, Rigobon 2022):\n\n1. Scope divergence (~38% of disagreement): Raters measure different things. One agency includes lobbying expenditures; another ignores it entirely.\n\n2. Measurement divergence (~56%): Even when measuring the same concept (e.g., 'labor practices'), agencies use different indicators, data sources, and methodologies.\n\n3. Weight divergence (~6%): Agencies assign different importance to the same categories (E vs. S vs. G).\n\n`mermaid\ngraph TD\n A[\"Same Company\"] --> B[\"Agency Alpha
Rating: AA\"]\n A --> C[\"Agency Beta
Rating: BB\"]\n A --> D[\"Agency Gamma
Rating: A\"]\n B --> E[\"Scope: 35 indicators
Weights: E=40% S=35% G=25%\"]\n C --> F[\"Scope: 28 indicators
Weights: E=25% S=25% G=50%\"]\n D --> G[\"Scope: 42 indicators
Weights: E=33% S=33% G=33%\"]\n E --> H[\"High E score
drives AA\"]\n F --> I[\"Low G score
drives BB\"]\n G --> J[\"Balanced
drives A\"]\n`\n\nConcrete Example:\n\nFalconridge Industries (hypothetical mining company) rated by three agencies:\n\n| Category | Pinecrest Ratings | Arbor ESG | Silverline Analytics |\n|---|---|---|---|\n| Environmental | A (strong mine rehabilitation) | C (high absolute emissions) | B (emission intensity improving) |\n| Social | B (good safety record) | A (community investment) | B (supply chain labor unclear) |\n| Governance | C (dual-class shares) | B (diverse board) | D (related-party transactions) |\n| Overall | B+ | B | C+ |\n\nSame company, three materially different conclusions. A portfolio manager using only Pinecrest would allocate; one using Silverline would exclude.\n\nImplications for Risk Management:\n\n1. No single source of truth: Using one ESG rating is equivalent to using one analyst's credit opinion while ignoring all others.\n\n2. Consensus approach: Average across 3+ raters to reduce idiosyncratic methodology bias. But averaging may obscure genuinely important disagreements.\n\n3. Materiality-focused: Use industry-specific materiality frameworks (SASB/ISSB) to weight the ESG dimensions that matter most for each sector.\n\n4. Internal assessment: Sophisticated investors increasingly build proprietary ESG models rather than relying on external ratings.\n\n5. Regulatory implications: The EU ESG Ratings Regulation (expected 2025-2026) will require transparency on methodology, conflicts of interest, and data sources.\n\nResearch Finding:\nPortfolios constructed using different ESG ratings have dramatically different holdings and returns. Aggregate ESG scores explain less than 20% of stock return variation, undermining claims that ESG integration systematically improves risk-adjusted performance.\n\nStudy ESG integration challenges in our FRM Part II course.