中文总结《实证宏观经济学：贝叶斯多元时间序列方法》系统阐述了如何应用贝叶斯计量技术来估计、评估和预测包含多个内生变量的大型宏观经济模型。其核心在于应对传统宏观计量方法在面对复杂、高维且可能存在不完全识别问题时面临的挑战，通过引入先验信息来提高模型的估计效率和预测稳健性。
本书的核心方法论围绕贝叶斯向量自回归及其扩展模型展开。它详细解释了如何将研究者或政策制定者对经济运行规律的理解（例如关于参数符号或大小的合理信念）量化为统计上的“先验分布”，并将其与数据信息（似然函数）结合，得到更新后的“后验分布”作为分析基础。这种方法在处理变量众多而时间序列数据有限的宏观情境下具有显著优势。全书通过具体案例（如分析货币政策冲击、预测经济周期）演示了贝叶斯方法如何被用于结构识别、模型比较、条件预测以及政策情景模拟，为现代宏观经济实证研究提供了一个强大且灵活的工具箱。
English TranslationEmpirical Macroeconomics: A Bayesian Multivariate Time Series Approach systematically explains how to apply Bayesian econometric techniques to estimate, evaluate, and forecast large-scale macroeconomic models with multiple endogenous variables. Its core addresses the challenges faced by traditional macroeconometric methods when dealing with complex, high-dimensional, and potentially under-identified systems, by incorporating prior information to improve estimation efficiency and forecast robustness.
The core methodology centers on the Bayesian Vector Autoregression model and its extensions. It details how to quantify a researcher's or policymaker's understanding of economic mechanisms as a statistical "prior distribution," which is then combined with data to form the "posterior distribution" for analysis. This approach offers significant advantages in macroeconomic contexts with many variables but limited time-series data. Through concrete applications, the book demonstrates how Bayesian methods are used for structural identification, model comparison, conditional forecasting, and policy scenario analysis, providing a powerful and flexible toolkit for modern empirical macroeconomic research.
