Anticipatory Skill and Structural Biases of a Major Global Risk Assessment

The authors evaluate the World Economic Forum’s Global Risks Reports (GRRs) from 2006–2024 by comparing survey-based risk likelihoods with a multi-decadal database of national shocks, and by tracing how surveys are translated into the reports’ narratives. Across eight shock categories, they find limited and uneven prospective alignment with observed shock frequency — with only geopolitical conflicts demonstrating clear anticipatory skill — and alignment consistently stronger for high-income than for lower-income regions, suggesting a potential regional bias. A thematic analysis reveals a transcription bias, as report narratives selectively downplay themes linked to poverty, inequality, biodiversity, and conflict while foregrounding economic growth, simplification, and short-termism relative to the surveys they rest on; a linguistic analysis further shows the GRRs have become increasingly negative, regulatory, and difficult to read over time. The authors conclude that GRRs are informative as a barometer of elite concerns but cast doubt on their use as early-warning tools, and call for a standing systemic risk assessment process that is system-based, empirically grounded, transparent regarding uncertainties, and developed through pluralistic and accountable processes.