Hidden bias in widely used economic risk measures

According to Casper de Vries and co-authors Milian Bachem and Lerby Ergun, these factors introduce a bias that affects the left and right tails of a distribution in opposite ways, creating the illusion of time-varying risk even when the underlying data-generating process remains stable. As a result, widely used conclusions about financial instability or economic extremes may be misleading.

Findings have broad implications for economics and finance

Importantly, the authors demonstrate that this bias is not only significant but also correctable. By exploiting the opposing effects in the two tails, they develop simple, practical methods that substantially reduce the distortion without requiring prior knowledge of the underlying factors. When applied to U.S. stock return data, the corrected measures provide more stable and accurate estimates of extreme risk and significantly improve the explanatory power of financial models.

The findings have broad implications for economics and finance. They suggest that past studies may have overstated changes in risk and highlight the importance of accounting for hidden factors when analyzing extreme outcomes. The authors emphasise that future research and risk measurement practices should explicitly consider these biases to avoid incorrect inferences and improve decision-making.