A0886
Title: Automatic slicing of connectivity networks for enhanced analysis of volatility spillovers
Authors: Heng Xiong - Wuhan University (China) [presenting]
Abstract: Effective portfolio management and hedging strategies hinge on accurately understanding the transmission of volatility across interconnected markets. A novel method is proposed that integrates nonparametric changepoint detection to automatically segment interconnectedness networks. Building upon this, the Diebold and Yilmaz (DY) framework is adapted to analyze these segments through return dynamics and EGARCH-filtered volatility. Unlike conventional methods, the sliced DY model significantly enhances the identification of shifts in market connectivity and minimizes divergence in conclusions across multiple measurement periods. To rigorously test this new approach, daily price data is employed from China's crude oil and agricultural futures markets. The empirical results robustly demonstrate that the proposed method successfully mitigates the issue of inconsistent findings across time metrics and enables a more nuanced examination of structural shifts.
