Assessing Industry Harm: OECD Reveals Risk Hotspots and Urgent Need for Better Data

The OECD’s cross-sector review finds strong global agreement that extractive industries, agriculture, and certain manufacturing sectors cause the most severe environmental, labour and human rights harms, while education, non-profits and real estate show consistently low impact. It also highlights major data gaps and inconsistencies for many mid-range sectors, underscoring the need for better, standardised information to guide responsible business conduct.

An extensive OECD study, built on research from KPMG, UNEP FI, GRI, S&P Global, Sustainalytics, RepRisk, the ILO, the USDOL, SBTN, and the Finance for Biodiversity Foundation, offers one of the clearest cross-sector portraits of environmental and social impacts to date. By merging these datasets with an expert survey of 133 practitioners, the OECD attempts to answer a deceptively simple question: Which sectors of the global economy are most closely linked to the world’s deepest social and environmental harms? What emerges is a story of strong agreement at the extremes, glaring gaps in the middle, and an urgent need for more consistent global data.

Extractives and Agriculture Dominate the High-Impact Landscape

Across nearly all sources, extractive industries, mining, coal, and oil and gas, top the charts for negative impacts. RepRisk controversy data shows mining leading human rights and environmental incidents, while expert respondents overwhelmingly link these industries to pollution, violent land conflicts, ecosystem collapse, and entrenched corruption. Agriculture, aquaculture, and fishing follow closely, reflecting widespread evidence of forced and child labour in USDOL findings, high fatality rates in ILO occupational injury records, and massive biodiversity and land-use disruption highlighted in SBTN scores. Certain manufacturing subsectors, especially textiles and apparel, chemicals, and food and beverages, recur as hotspots for labour exploitation, waste mismanagement, pollution, and consumer harm. Visuals in the report consistently cluster these sectors at the high-risk end of nearly every impact category.

Sectors with Narrow but Intense Risks

The analysis shows that not all sectors cause broad-spectrum harm; some exhibit tightly focused but severe risk profiles. Financial services, banking, insurance, and capital markets are repeatedly linked to corruption, deceptive marketing, consumer fraud, and tax avoidance. Software, media, and communication sectors feature prominently in technology-related risks, particularly privacy violations, personal data misuse, and intellectual property theft. These sectors barely register on environmental or labour indicators, yet dominate technology and consumer-related harms. Construction stands out as an industry straddling multiple pressures: one of the highest frequencies of occupational injuries in ILO data, a major target for corruption allegations in RepRisk datasets, and a sector experts consistently associate with dangerous and unfair working conditions.

Where the World Seems Least at Risk

Throughout the datasets, there is unusually strong agreement on sectors that appear comparatively low-impact. Education services, non-profit organisations, real estate, and certain light manufacturing activities rarely appear in litigation databases, controversy monitoring tools, labour statistics, or expert perceptions. Scatterplots of normalised scores in the report show these sectors clustering near zero across environmental, corruption, and human-rights categories. While no sector is risk-free, these consistently low associations suggest limited exposure to the kinds of harms covered under the OECD Guidelines. Even here, however, the report warns against complacency: low visibility could also reflect poor reporting or weak data coverage.

The Troubling Middle: Data Gaps and Conflicting Signals

The most challenging territory is the wide “middle band” of sectors where findings diverge sharply depending on the dataset. Utilities, for instance, appear extremely high-risk for human rights controversies in RepRisk data, yet score far lower in expert perceptions. Pharmaceuticals show the greatest variability of all sectors, sometimes ranking high for consumer risk and waste issues, other times showing moderate or low exposure, reflecting conflicting methods, incomplete environmental data, and inconsistent value-chain boundaries. Electronics, transportation infrastructure, renewable energy, and various consumer-facing industries also fall into this zone of contradiction. This inconsistency stems from incompatible classification systems, mixed materiality lenses, and the uneven ability of datasets to capture indirect or supply-chain impacts. The OECD’s attempt to harmonise these systems into a unified 40-sector framework underscores how deeply fragmented global sustainability metrics remain.

In its final reflections, the OECD emphasises that despite strong signals about the most harmful sectors, extractives, agriculture, and certain types of manufacturing, the wider picture is marked by uneven data, incompatible classifications, and blind spots in supply-chain visibility. Sector comparisons can help businesses and policymakers prioritise risks, but they cannot replace company-specific due diligence. The report calls for more standardised and transparent global data to ensure that firms, especially SMEs, can properly identify and address the environmental and social harms embedded in their operations and value chains.