Seymon Malamud (Ecole Polytechnique Federale de Lausanne): APT or “AIPT”? The Surprising Dominance of Large Factor Models

The Stanford AFTLab invites you to the AI & Big Data in Finance Research Forum (ABFR) webinar:

The webinar is on October 31, 9-10am Pacific Time (12-1pm ET)

Presenter: Seymon Malamud (Ecole Polytechnique Federale de Lausanne)

Discussant: Paul Schneider (University of Lugano)

Zoom webinar link:https://stanford.zoom.us/j/97599565536?pwd=l7aVKGz3bXD2h7xgeHftsp2BGcGhle.1

Webinar ID: 975 9956 5536

Passcode: 676284

For more information, please visit our website:https://www.abfr-forum.org

To stay up to date please join our mailing list:https://groups.google.com/u/0/g/abfr-forum

Title:  APT or “AIPT”? The Surprising Dominance of Large Factor Models 

Authors: Antoine Didisheim (Melbourne), Shikun Ke (Yale), Bryan Kelly (Yale), and Semyon Malamud (EPFL)

Abstract: We introduce artificial intelligence pricing theory (AIPT). In contrast with the APT’s foundational assumption of a low dimensional factor structure in returns, the AIPT conjectures that returns are driven by a large number of factors. We first verify this conjecture empirically and show that nonlinear models with an exorbitant number of factors (many more than the number of training observations or base assets) are far more successful in describing the out-of-sample behavior of asset returns than simpler standard models. We then theoretically characterize the behavior of large factor pricing models, from which we show that the AIPT’s “many factors” conjecture faithfully explains our empirical findings, while the APT’s “few factors” conjecture is contradicted by the data.

Bio of speaker: Semyon Malamud is an Associate Professor of Finance at the Ecole Polytechnique Federale de Lausanne and a Senior Chair at the Swiss Finance Institute. Professor Malamud is a regular speaker at leading academic conferences worldwide, and his papers have been published in the top journals in finance and economics. He is an Associate Editor for the Journal of Finance, the Journal of Financial Econometrics, and the Journal of Behavioral and Experimental Finance. Professor Malamud focuses on predicting asset prices and market returns. While the industry is experiencing a boom in the adoption of machine learning techniques to improve portfolio construction, little is known about the underlying theoretical processes. His theoretical work shows that simple models severely understate return predictability compared to "complex" models. In other words, the performance of machine learning portfolios can be improved by further pushing the level of model complexity. Empirical results show that US equity market data is remarkably well aligned with complex frameworks. These findings are not a license to add every arbitrary predictor one encounters to a machine learning–based model, but do suggest that the finance profession should focus on rich nonlinear models that include plausibly relevant predictors. He obtained his PhD in mathematics at the Swiss Federal Institute of Technology Zurich, where he previously also was an Assistant Professor before moving to EPFL.  

Bio of discussant: Paul Schneider is a Full Professor at the University of Lugano and a Senior Chair at the Swiss Finance Institute. His research interests include financial econometrics, statistical methods, and asset pricing. He is interested in finding structure in financial markets and his research focuses is on methods to recover information that is not directly observable with the most innocuous assumptions possible. His work has been published in the Annals of Statistics, Journal of Econometrics, Management Science, and the Journal of Finance among other international journals. He is an Associate Editor of the Journal of Financial Econometrics. He graduated in 2006 from the University of Vienna with a PhD in finance. After a post-doctoral position in Vienna he joined the University of Warwick in 2008 as Assistant Professor. Since 2012 he has been a professor in quantitative methods at the Faculty of Economics at the University of Lugano.