中国经济周期的混频数据测度及实时分析

Measuring China's Business Cycle with Mixed-frequency Data and Its Real Time Analysis

现代宏观经济研究表明,周期波动体现了月度、季度及年度等各种宏观经济指标的协同性变动。考虑到GDP等季度指标在宏观经济分析中的重要地位,本文构建了一种能够综合利用我国季度数据和月度数据的经济周期计量模型,即混频数据区制转移动态因子模型。通过对我国季度GDP同比增长率和五个月度一致指标进行实证建模和估计,本文不仅识别了我国1992年至2011年间的经济周期变化,而且获得了可以描述我国经济运行状况的一致指数。特别地,本文通过收集宏观经济实时数据,进一步从实时分析的角度考察了该模型在我国经济周期测度(经济转折点识别和测定)上的可靠性和时效性,从而验证了该模型在我国的适用性。

Modern macroeconomic studies show that economic fluctuations behave business cycle comovement across the monthly, quarterly and annually economic indicators. Taking into account the key role of quarterly indicators such as GDP in macroeconomic analysis, this paper introduces an econometric model （ called mixed-frequency Markov switching dynamic factor model） that could combine both monthly and quarterly indicators to model the business cycle. Though our empirical analysis on China＇s year-on-year quarterly GDP growth rate and five monthly coincident indicators, we can not only identify business cycle phases from 1992 to 2011, but also extract a new coincident index to describe China＇s economic conditions. Furthermore, based on the real time data set we collected, this paper examines the reliability and timeliness of this model in identifying business cycle turning points and thus verifies its applicability for China.