统计与数据科学学院知识库

从08年全球金融危机开始，金融系统的波动对宏观经济的影响已经变得越来越不可忽视，监测和防范系统性金融风险显得愈发重要。而金融状况指数（FCI）是一种衡量金融市场整体变化的综合指标，可以作为货币政策和宏观经济的指示器及测度指标。因此,中国FCI构建及其动态特征分析对我国防范经济金融风险至关重要。

本文在已有研究的基础上基于贝叶斯估计方法构造两种动态金融状况指数，并对其波动特征和权重差异以及金融状况指数同我国宏观经济之间的关联性进行了比较。本文的主要研究内容如下：

首先利用动态因子模型和引入贝叶斯估计方法的总需求缩减方程式及TVP-VAR模型分别构建了FCI，并分析和对比了指数波动特征以及动态权重变化。其次，通过格兰杰因果检验、跨期相关性分析和回归方程初步检验和对比金融状况指数的有效性。再次，利用频域分析方法对构建指数和宏观经济变量之间的周期关联性质进行测度和比较。最后，通过马尔可夫区制转换模型研究金融状况指数本身的变动特征并比较其中差异。

研究结果表明：第一，通过不同方法构建出的金融状况指数总体趋势差异不大，在金融变量对金融市场的动态影响刻画上也具有一定的共性，但是各自有所侧重。第二，两种金融状况指数都能在短期有效的预测通货膨胀，但预测能力有强弱之分。第三，金融状况指数和宏观经济主周期长度相近，但两种金融状况指数对产出、宏观经济和通货膨胀呈现出的周期联动关系有强弱侧重的差异，而领先周期差距并不大。第四，金融状况指数呈现出一定的周期波动变化，金融状况指数各个状态都具有一定的稳定性，但两种金融状况指数对金融市场扩张状态的判定和识别存在一定差异。

综上所述，不同方法构建的金融状况指数在波动特征和权重差异，以及同我国宏观经济之间的关联性方面各自有所侧重，而针对不同金融状况指数表现出的特征属性应确定其主要应用方向和应用范围。由此可知，货币当局应利用多种方法定期编制金融状况指数，根据不同指数特征差异针对性的调整货币政策。

Since the 2008 global financial crisis triggered by the US subprime mortgage crisis, the impact of the fluctuation of the financial system on the macroeconomy has become more and more important. It is more and more important to supervise and to be on alert for systemic financial risks. The Financial Condition Index (FCI) is a comprehensive index for measuring the overall changes of the financial market, which was an indicator and measurement index of monetary policy and macro-economy. Therefore, the construction of China's FCI and its dynamic characteristics analysis are very important for China to prevent economic and financial risks.

Based on existing research, this paper constructs two dynamic Financial Condition Indexes based on the Bayesian estimation method; and compares their fluctuation characteristics and weight differences, as well as the correlation between Financial Condition Index and China's macro-economy. The major contents of this paper are as follows:

Firstly, the FCI is constructed by using the dynamic factor model, the aggregate demand reduction equation with the Bayesian estimation method, and TVP-VAR model, and the fluctuation characteristics and dynamic weight changes of the index are analyzed and compared. Secondly, through the Granger causality test, intertemporal correlation analysis, and regression equation, the effectiveness of the financial condition index is preliminarily tested and compared. Thirdly, the frequency domain analysis method is used to measure and compare the correlation between Financial Condition Index and the macro-economy. Finally, the volatility characteristics of the financial condition index itself are studied and the differences are compared through the Markov regime-switching model.

The results of research show that: First, The overall trend of the Financial Condition Index constructed by different methods is not different, and it also has some commonalities in the description of the dynamic impact of financial variables on the financial market, but each has its emphasis. Second, both financial condition indexes can effectively predict inflation in the short term, but their prediction ability can be divided into strong and weak. Third, the Financial Condition Index and the macroeconomic main cycle are similar in length, but the two Financial Condition Indexes have different emphasis on the cyclical linkage relationship between output, macroeconomy, and inflation, while the gap between leading cycles is not large. Fourth, the financial condition index shows a certain periodic fluctuation. Each state of the financial condition index has a certain stability, but there are some differences in the judgment and identification of the expansion state of the financial market between the two Financial Condition Indexes.

To sum up, the Financial Condition Index constructed by different methods focuses on the fluctuation characteristics, weight differences, and the correlation with China's macro-economy, and its main application direction and scope should be determined according to the characteristic attributes of different financial condition indexes. It can be seen that the monetary authority should use a variety of methods to regularly prepare the Financial Condition Index and adjust the monetary policy according to the characteristics of different indexes.