Environmental, Social, and Governance (ESG) metrics, aggregated in rating products, increasingly inform a wide range of business and investment decisions. For policy makers, investors, business and other stakeholders, having sufficient levels of transparency and quality with respect to the ESG metrics used in these products is key to ensuring clarity, consistency, and accountability when it comes to measuring sustainability performance of business. For investors, this means making informed investment decisions and for policy makers being able to monitor and evaluate the impact of sustainable finance-related policies. Against this background, this report aims to assess the scope and characteristics of over 2 000 ESG metrics from eight major ESG rating products. The analysis helped identify four key findings as presented below. Metric scope: significant imbalances and gaps across ESG topics More nascent or less standardised ESG issues typically lack comprehensive and granular metrics compared to more established topics. For instance, over 20 different metrics are used on average to measure performance related to topics such as corporate governance, business ethics and environmental management, compared to less than five metrics for topics such as biodiversity, business resilience, and community relations. In some cases, certain topics are entirely omitted from ESG rating products, including human rights and corruption. While it cannot be assumed that a higher number of metrics leads to better measurement, an extremely limited number of metrics associated with a topic may infer that topical impacts, risks, and opportunities are not being captured in a meaningful and comprehensive way. Metric comparability: Considerable divergences in measurement approaches across products Significant divergences exist when comparing the scope of metrics for the same topic across rating products. For instance, one rating product uses 28 times more metrics to measure Corporate Governance performance compared to another. The range varies from 1 to 47 metrics to measure corporate GHG Emissions, and from 4 to 113 metrics to gauge a company’s corporate governance. High variations in the number of metrics available per topic across rating products usually reflect distinct methodological approaches, divergent levels of granularity applied, and likely disagreement as to how performance ought to be measured. Metric characteristics: ESG performance largely measured by focusing on business’ effort rather than effect ESG rating products rely primarily on input-based metrics (68%). These metrics capture self-reported policies and activities put in place to address potential and actual ESG impacts, risks, and opportunities. Meanwhile, a third of the metrics rely primarily on output-based metrics, focusing on the outcomes of these policies and activities. The reliance on input-based metrics could incentivise “tick-boxing” approaches over actual risk prevention and mitigation actions. It may also benefit large companies over SMEs, as multinational enterprises may have more resources to adopt, implement and disclose measures underpinned by such metrics. Moreover, ESG performance is predominantly assessed using qualitative metrics (72%). Noticeably, input based metrics account for the vast majority of qualitative metrics. These metrics may not always provide a reliable proxy of a company’s ESG performance but rather infer ESG performance based on the existence of policies and measures to manage impacts, risks and opportunities related to that topic, irrespective of the actual effectiveness of such measures. Conversely, only 17% of all metrics are quantitative output-based metrics. For policy makers, the potential disconnect between the proxy metric and the performance measurement can also have implications with regard to assessing the effectiveness of public policies. Lastly, there seems to be a positive correlation between low shares of quantitative data and low numbers of metrics per topic. For instance, biodiversity, climate resilience, taxation, and competition are among the topics with the lowest shares of quantitative data and the lowest number of metrics overall, further suggesting that assessment of performance against certain topics may not be sufficient. Looking forward: metrics are insufficient to assess observance of OECD standards on responsible business conduct OECD instruments on responsible business conduct promote risk-based due diligence, including the identification and prioritisation of adverse impacts. In contrast, ESG rating products tend to measure how companies manage impacts, risks, and opportunities with respect to a specific topic—not across topics—irrespective of their interlinkages and interdependencies. Less than 5% of input-based metrics could be associated with explicit risk-based due diligence measures and steps without being associated with one single topic. This siloed and topical structure is also at odds with recent sustainability-related standards structures (e.g. ESRS and ISSB), creating potential challenges for investors wishing to leverage ESG metrics to assess the quality and effectiveness of companies’ due diligence across sustainability issues. Moreover, most ESG rating products assess observance or “violations” of the OECD Guidelines through controversy-related metrics as a proxy. These metrics usually look at the existence and prevalence of controversies in a company’s operations and/or supply chains, rather than evaluating a company’s due diligence efforts and effectiveness in mitigating sustainability impacts. 15% of all metrics could be broadly identified as ‘controversy-based’. Finally, measurement of ESG performance beyond an entity direct operation is limited, including measurement of how businesses identify, prevent, mitigate and account for adverse impacts in their business relationships and global supply chains. Only 7% of all metrics could be associated with supply chain risk management metrics across topics and products.