Understanding Extreme Market Behavior: The Efficient Tail Hypothesis
About
A new study by researchers Junshu Jiang, James Richards, Raphaël Huser, and David Bolin introduces the Efficient Tail Hypothesis (ETH), an analogue of the Efficient Market Hypothesis focused on extreme events in financial markets. Drawing on extreme value theory, the team developed a novel statistical measure to evaluate whether asset markets remain informationally efficient even during rare, extreme fluctuations. Their findings not only challenge conventional wisdom on market efficiency but also open the door to identifying potential opportunities (and risks) that emerge during market extremes.
Financial economists have long relied on the Efficient Market Hypothesis (EMH), which posits that asset prices instantly reflect all available information. In practice, however, markets can behave erratically during financial crises, suggesting that efficiency may falter under extreme conditions. Traditional models for extreme market events are often limited. This new study addresses these gaps by constructing regular variation models on the entirety of multidimensional space and introducing balanced regular variation, a framework where both upper and lower tails have unit scale, allowing extremal dependence to be modeled simultaneously for both directions of extreme movements.
At the heart of the research is a Directional Tail Dependence (DTD) metric, which quantifies asymmetries in how extreme losses or gains in one asset correspond to extremes in another. Using DTD, the authors define the Efficient Tail Hypothesis (ETH) as a testable version of EMH for the tails of distribution. Under ETH, no disproportionate information leak or arbitrage opportunity exists in the market’s most extreme ups or downs. The study provides theoretical guarantees for the DTD estimator and introduces a permutation-based statistical test (with accompanying visualization tools) to assess whether real markets satisfy the ETH.
The practical value of the ETH was demonstrated through an extensive empirical analysis of China’s futures market, where the hypothesis was statistically rejected. The study revealed systematic tail inefficiencies. Negative extreme depreciation events proved more influential than positive extremes, and co-directional lead-lag effects across assets indicated that the market tends to be under-active during extreme situations. These findings contradict the assumption of efficient information processing during crisis periods and suggest potential profitable strategies during market turmoil. To promote further research in this area, the team has open-sourced their extensive high-frequency dataset from China’s derivatives market at GitHub, continuously updated from multiple exchanges, enabling other researchers and practitioners to explore market microstructure and develop tail-risk strategies.
REFERENCE:
Jiang, J., Richards, J., Huser, R., & Bolin, D. (2025). The Efficient Tail Hypothesis: An Extreme Value Perspective on Market Efficiency. Journal of Business & Economic Statistics, in press.
