Dynamics of hierarchical clustering in stocks market during financial crises

Abstract

We examine the behaviors of stocks in the Stock Exchange of Thailand (SET) during financial crises from 2008 to 2020 by using a variety of indicators such as average entropy, average correlation, and average Fisher information distance. All of these indicators increased dramatically during a financial crisis, which agreed with the Ulcer index indicating the root mean square of the percentage price drawdown. We investigate the dynamics of the hierarchical tree structure of 37 stocks in SET, using the correlation and the Fisher information distance between stocks in different time windows. The Fisher information distance is more robust, can provide earlier information, and yields a much-slower decay of signal-to-noise ratio as the averaging window size increases. As the hierarchical clustering evolves during a period of a financial crisis, its structure has lower tendency than usual to change the clusters, as evidenced by the low variation of information and anticorrelation of stocks. The dynamic hierarchical using the Fisher information distance gives less-variant clustering and can shed some insight on stock behaviors during a financial crisis.