Category Learning, Trade Volume, and Artificial Intelligence

Viernes 1/11, 13.15h

Seminario de Negocios | Paula Margaretic
Abstract:
This paper presents a noisy rational expectations model in which attention-constrained investors allocate their attention between asset-specific fundamentals and categorical factors influencing groups of assets. The model predicts that more attention is directed toward larger categories, which leads to increased trade volumes in assets belonging to those categories. We empirically validate these predictions by analyzing data from the social media platform X (formerly Twitter), focusing on how users mention and co-mention assets over time. We construct a dynamic network of co-mentions and use community detection algorithms to produce time-varying measures of asset-specific and categorical attention. We then explore the impact of artificial intelligence (AI) on investors’ behavior, showing that AI enhances categorical attention while also increasing selectivity in community formation. Finally, we find that AI negatively influences trade volume and positively correlates with trade volume volatility. Our results have significant implications for understanding the role of AI in reshaping market dynamics and investor behavior.

Bio:
PhD in Finance from the University of Toulouse (2012), Paula is currently working at the University of Chile as a researcher and assistant professor of finance. Previously, she worked at the University Adolfo Ibañez (Chile), San Andrés University in Argentina, Central Bank of Chile, Airbus France, Banco Santander in Argentina and Universidad Argentina de la Empresa (UADE) in Argentina, among others.

Lugar: Sala principal de Rectorado - 3.er piso Edificio Alcorta | Campus Di Tella: Av. Figueroa Alcorta 7350, Ciudad de Buenos Aires.
Contacto: Mariana Cunillé