Anticipatory skill and structural biases of a major global risk assessment

Louis Delannoy, Mélis Busson, Peter Søgaard Jørgensen

Calls to navigate the global polycrisis have multiplied, yet one of the most visible systemic risk assessments—the World Economic Forum’s Global Risks Reports (GRRs)—has not been comprehensively evaluated. We analyse all GRRs from 2006–2024 by reconstructing and comparing survey-based risk likelihoods with a multi-decadal database of national shocks, and by tracing how surveys are translated into the reports’ narratives. Despite increasing concerns for environmental risks, we find that GRRs show limited and uneven prospective alignment with observed shock frequency across categories, and that only geopolitical conflicts demonstrate clear anticipatory skill. Alignment is also consistently stronger for high-income than for lower-income regions, suggesting a potential regional bias. In parallel, we document a transcription bias, as the reports selectively and significantly downplay themes linked to poverty, inequality, biodiversity, and conflict, while foregrounding economic growth, simplification, and weaker understandings of sustainability relative to the surveys they rest on. We also note a framing bias in the way GRRs increasingly portray risks as negative, complex and regulatory challenges, hinting at a shift in the target audience over the years towards a technical policy and business one. Our findings suggest that GRRs are informative as a barometer of elite concerns, but cast doubt on their use as early-warning tools. We conclude by proposing that global risk governance uses assessments that are system-based, grounded in empirical data, transparent regarding uncertainties, and developed through pluralistic and accountable processes.