Investor behavior and multiscale cross-correlations: Unveiling regime shifts in global financial markets

Marina Dolfin,George Kapetanios,Leone Leonida,Jose De Leon Miranda

2024-08-30

Abstract:We propose an algorithm to capture emergent patterns in the cross-correlations of financial markets, highlighting regime changes on a global scale. In our approach, financial markets are viewed as complex adaptive systems, and multiscale properties and cross-correlations are considered, particularly during stress conditions such as the COVID-19 pandemic, the invasion of Ukraine by Russia in 2022, and Brexit. We investigate whether significant disruptions reflect an imbalance in investment horizons among investors, and we propose a measure based on this imbalance to depict the impact on global financial markets. The detrended cross-correlation cost (DCCC), which is derived from detrended cross-correlation analysis, uses cross-correlations at different timescales to capture variations in investment horizons amid financial uncertainties. Our algorithm, which combines DCCC analysis and the minimum-spanning-tree filtering approach, tracks system interconnectedness and investor imbalances. We tested the DCCC indicator using daily price series of G7, Russian, and Chinese markets over the past decade and found that it increases sharply during ``crash'' periods compared to ``business as usual'' periods. Our empirical results confirm that short-term investment horizons dominate during financial instabilities; this validates our hypothesis and indicates that the DCCC can serve as a leading indicator of shifts in financial-market regimes.

General Economics

What problem does this paper attempt to address?

The paper aims to address the issue of pattern recognition in cross-scale cross-correlations in financial markets, with a particular focus on changes in these patterns during periods of market instability (such as events like the COVID-19 pandemic, the Russia-Ukraine conflict, and Brexit). Specifically, the research objectives include:

1. **Identifying patterns of market instability**: By analyzing nonlinear cross-correlations across different time scales to identify financial instability patterns that may lead to regime shifts in the market.

2. **Proposing new indicators**: Introducing the Detrended Cross-Correlation Cost (DCCC) to capture changes in investment horizons during periods of financial uncertainty.

3. **Validating hypotheses**: Verifying whether short-term investors dominate during periods of financial instability and reflecting this impact through changes in the DCCC indicator.

4. **Monitoring market interconnectedness and investor imbalance**: Combining DCCC analysis with the minimum spanning tree filtering method to track system interconnectedness and investor imbalance.

5. **Applying real data**: Using daily closing price data from the past 10 years for the G7 countries, China, and Russia to conduct empirical tests, demonstrating the applicability of this method to real-world financial data.

Overall, the study aims to identify regime shifts in financial markets by monitoring changes in market network structures across different time scales, providing investors with effective risk management and asset allocation strategies.