Financial markets are interconnected and fragile making them vulnerable to systemic contagion, and measuring this risk is crucial for regulatory responsiveness. This study introduces a new set of measures for systemic risk using a copula-based (CB) estimation method with a focus on U.S. Bank Holding Companies. Unlike most of the prevailing systemic risk measures, CB methodology relies on balance sheet data, instead of market price data, which makes it globally applicable. We compared CB measures with three existing measures of systemic risk that rely on market data and find that CB measures provide competitive results, in both the short and medium term, for systemic risk forecasting. The forecasting evaluation shows that CB measures perform consistently better than historical unconditional quantile of macroeconomic indicators. By using out-of-sample predictive quantile regression, we ascertain that CB systemic risk measures can forecast the 10th and 20th percentile movements of different macroeconomic indicators up to 6 quarters in advance. Moreover, systemic risk measures, existing as well as CB, are better predictors of the 20th percentile shocks to sector-specific indicator and 10th percentile shocks to broader macroeconomic indicators.

中文翻译：

---

系统性风险和宏观经济预测：一种全球适用的基于 copula 的方法

金融市场相互关联且脆弱，容易受到系统性传染，衡量这种风险对于监管响应至关重要。本研究介绍了一套新的系统性风险措施，使用基于联结 (CB) 的估计方法，重点是美国银行控股公司。与大多数流行的系统性风险措施不同，CB 方法依赖于资产负债表数据，而不是市场价格数据，这使其在全球范围内适用。我们将 CB 指标与依赖于市场数据的三种现有系统性风险指标进行了比较，发现 CB 指标在短期和中期为系统性风险预测提供了有竞争力的结果。预测评估表明，CB 指标的表现始终优于宏观经济指标的历史无条件分位数。通过使用样本外预测分位数回归，我们确定 CB 系统性风险度量可以提前 6 个季度预测不同宏观经济指标的第 10 个和第 20 个百分点的变动。此外，现有的系统性风险指标以及 CB 是对特定行业指标的第 20 个百分点冲击和对更广泛宏观经济指标的第 10 个百分点冲击的更好预测指标。