Speaker: Roberto Reno, ESSEC Business School (https://sites.google.com/essec.edu/reno/)

Title: A seismometer for Ito semimartingales

We propose a novel statistical method for detecting explosive behavior in It.o semimartingales with unbounded drift and volatility. We show that the high-frequency asymptotic behavior of the localized autocovariance reveals both the intensity and the source of the explosion. Building on this insight, we develop a testing procedure that classifies explosion regimes and remains valid under irregular sampling and measurement noise. In particular, strongly negative values of the test statistic provid evidence of volatility-driven explosions. Realistic simulations of financial data demonstrate the effectiveness of our methodology. We apply the test to show that financial markets were not stabilized by the market-wide circuit breakers in March 2020, which were activated during the turmoil triggered by COVID-19.

Host: Kim Christensen

Organisers: Stefán Guðmundsson and Mikkel Sølvsten