标题： 投资者行为与多尺度交叉相关性：揭示全球金融市场中的制度转变 显示英文标题

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

摘要： 我们提出一种算法来捕捉金融市场交叉相关性中的新兴模式，并突出全球范围内的制度变化。 在我们的方法中，金融市场被视为复杂适应系统，考虑了多尺度特性和交叉相关性，特别是在压力状况下，如新冠疫情、2022年俄罗斯对乌克兰的入侵和英国脱欧。 我们研究重大中断是否反映了投资者投资期限之间的失衡，并提出一种基于这种失衡的度量来描述对全球金融市场的影响。 去趋势交叉相关成本（DCCC），它源自去趋势交叉相关分析，利用不同时空尺度的交叉相关性来捕捉金融不确定性中的投资期限变化。 我们的算法结合了DCCC分析和最小生成树过滤方法，跟踪系统相互关联性和投资者失衡。 我们使用过去十年G7、俄罗斯和中国市场每日价格序列测试了DCCC指标，发现与“正常”时期相比，在“崩溃”时期该指标急剧上升。 我们的实证结果证实，在金融不稳定期间短期投资期限占主导地位；这验证了我们的假设，并表明DCCC可以作为金融市场制度转变的领先指标。 显示英文摘要

摘要： 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.