Firmographica: Knowledge Graph and AI-based Framework for Short-Selling Risk Assessment

Abstract

This study investigates how corporate ownership structures affect short selling in publicly traded banking firms by integrating knowledge graph methodologies with regression analysis. By combining graph-based centrality metrics with traditional indicators like firm size and ownership concentration, we systematically identify and rank the factors that influence short-selling activity. Our results indicate that firm size, the degree of ownership concentration, and PageRank centrality are consistently significant predictors of both the level and intensity of short-selling positions. Additionally, insider trading activity is shown to be a critical determinant of short-selling volatility, suggesting that internal market behaviors provide unique predictive value beyond standard financial metrics. The findings underscore the broader potential for graph analytics and machine learning approaches to enhance financial risk modeling and market surveillance. By providing a richer, network-driven perspective, this research contributes to a better understanding of market dynamics, supports the development of more robust governance frameworks, and informs both regulatory policy and investment strategies aimed at promoting transparency and stability in the financial sector.