Distributional Vector Autoregression: Eliciting Macro and Financial Dependence
提出分布向量自回归模型，无需参数分布假设即可建模多元时间序列，并开发分布脉冲响应函数以揭示动态异质性。应用于美国数据，该模型在预测和宏观动态分析中表现优于现有方法。
ABSTRACT This paper extends the vector autoregression framework by introducing a flexible distributional regression that models multivariate time series without imposing restrictive parametric distribution assumptions. We develop a distributional impulse response function that captures the future effect of distributional disturbances within the system, providing a more detailed view of dynamic heterogeneity. We propose a straightforward estimation method and establish its asymptotic properties under weak dependence assumptions. Our model, in an application to U.S. economic data, exhibits strong forecasting capabilities compared to existing alternatives. By examining distributional interactions and monetary policy impacts, particularly during the Great Recession, we uncover complex macroeconomic dynamics.