Averaged static connectedness.

Most of the existing studies on the connectedness among economic policy uncertainties (EPUs) usually neglect the quantile and frequency domain perspectives. To address this limitation, this paper proposes a quantile time–frequency connectedness model to analyze the connectedness among EPUs by combining the quantile and frequency domain dimensions. First, the quantile-vector autoregressive model (QVAR(p)) is estimated and converted into the quantile-vector moving average representation (QVMA()). Next, the generalized prediction error variance decomposition (GFEVD) is computed, from which various types of time-domain connectedness metrics are calculated. Finally, the spectral decomposition method is used to compute frequency-domain connectedness metrics and establish a link between time- and frequency-domain metrics. The empirical results of this paper, based on the sample data of China and G7 countries, reveal several important findings. The EPU of the United States acts as a net transmitter of shocks in both the short and long term, whereas China functions as a net receiver of shocks. The total connectedness index (TCI) demonstrates significant heterogeneity, with its dynamics primarily driven by short-term rather than long-term components. Additionally, connectedness shows substantial improvement under extreme conditions.

现有关于经济政策不确定性（Economic Policy Uncertainty，EPU）联动性的多数研究往往忽略了分位数与频域视角。为弥补这一研究局限，本文提出分位数时频联动性模型，结合分位数与频域维度分析EPU间的联动性。首先，估计分位数向量自回归模型（Quantile-Vector Autoregressive Model，QVAR(p)）并转换为分位数向量移动平均表征（Quantile-Vector Moving Average Representation，QVMA）；其次，计算广义预测误差方差分解（Generalized Prediction Error Variance Decomposition，GFEVD），据此得到各类时域联动性测度指标；最后，借助谱分解方法计算频域联动性测度指标，并建立时域与频域指标间的关联。本文基于中国与G7国家的样本数据开展实证研究，得到若干重要结论：美国的EPU在短期与长期均为冲击净传导方，而中国则为冲击净接收方；总联动性指数（Total Connectedness Index，TCI）呈现显著异质性，其动态变化主要由短期成分而非长期成分驱动；此外，极端情形下的联动性水平会出现显著提升。