ESG Company Scores

Overview

Veridion provides a high-fidelity, representative sample of a 134 million company universe that covers the full spectrum of existing public and private entities, especially bridging critical data gaps within the private sector and SME landscape. The ESG Company Scores dataset extends this intelligence with comprehensive environmental, social, and governance performance assessments, enabling research into corporate sustainability practices across both public and private firms.

Data Description

The Veridion ESG Company Scores dataset is a curated sample of 5,000 companies drawn from Veridion's broader universe of 134M+ firms, providing comprehensive environmental, social, and governance (ESG) performance assessments across multiple industries and geographies. The number of companies in the dataset is expected to grow as additional companies will be passed through the Veridion ESG evaluation pipelines. The dataset includes an overall ESG score (0–100), individual Environmental, Social, and Governance pillar scores, and 26 granular risk-criteria scores aligned with the UNEP FI Principles for Sustainable Insurance taxonomy, each accompanied by a natural-language justification. Scores are generated by proprietary models using structured Veridion company data, publicly available ESG disclosures (official commitments and news), and industry context, with the methodology prioritizing demonstrated real-world actions over stated commitments. Industry benchmarking is included via NAICS-based industry averages and quartile rankings, enabling relative performance assessment. Companies without sufficient ESG-related data are excluded to maintain scoring credibility. The sample is designed to illustrate the depth and structure of Veridion's ESG scoring methodology and is particularly suited for exploratory research, methodological evaluation, and pilot studies on ESG performance among both public and private companies.

Observation Level: Veridion ID (unique ID per company) → Company-level records with the corresponding ESG scores and justifications.

Historical Coverage: Datasets provide cross-sectional snapshots rather than longitudinal panels, as the collection methodology prioritizes accuracy of current state over historical preservation. Researchers should treat each release as an independent cross-section of the global firm population. This is especially relevant for the ESG score as Veridion scans public sources for the most recent evidence, such as news articles, to ensure that the ESG scores are a representation of the actual ESG performance of a company to date.

Coverage

The ESG Company Scores dataset is a curated sample of 5,000 companies (expected to grow) drawn from Veridion's broader universe of 134M+ firms, spanning multiple industries and geographies. Companies without sufficient ESG-relevant data are excluded rather than assigned default scores to maintain scoring credibility.

Methodology

The ESG scores are generated by combining Veridion's proprietary structured company data with publicly available ESG disclosures, news articles, and regulatory filings. Each company is first matched to its Veridion profile through the entity resolution engine, then enriched with detailed firmographic and operational attributes. Proprietary AI-powered large language models analyze this combined input using optimized, independent prompts for each score dimension, ensuring consistency and transparency. Scores prioritize evidence of real-world ESG actions over stated commitments, and industry context is incorporated via NAICS-based sectoral benchmarks. Companies without sufficient ESG-relevant data are excluded rather than assigned default scores. The taxonomy of 26 risk criteria was developed in collaboration with insurance industry practitioners and aligned with the UNEP FI Principles for Sustainable Insurance framework.