Mathematical Finance & Financial Data Science Seminar

Excess out-of-sample risk, fleeting modes and Random Matrix Theory

Speaker: Jean-Philippe Bouchaud, Capital Fund Management

Location: Online Zoom access provided to registrants

Date: Tuesday, April 25, 2023, 5:30 p.m.


Using Random Matrix Theory, we propose a universal and versatile tools to reveal, quantify and characterize non-stationarity effects in the correlation structure of large asset universes. We discuss in particular the existence of “fleeting modes”, i.e. portfolios that carry statistically significant excess risk, signalling ex-post a change in the correlation structure in the underlying asset space. Our proposed tests are, somewhat miraculously,  independent of the “true” (but unknown) underlying correlation structure.

Speaker Bio:

Jean-Philippe Bouchaud is co-founder, chairman and chief scientist of Capital Fund Management and adjunct Professor at Ecole Normale Supérieure. Jean-Philippe Bouchaud has received numerous national and international honours: He is elected as member of the French Academy of Sciences. In 2020 he held the Bettencourt Innovation Chair at Collège de France.  In 2017 he was named Quant of the Year by Risk Magazine.



Presentation Slides by J.-Ph. Bouchaud


My Beautiful Laundrette: Cleaning Correlation Matrices for Portfolio Optimization

Conditional Correlations and Principal Regression analysis for Futures

Principal regression analysis and the index leverage effect

Overlaps between eigenvectors of correlated random matrices