大维不可观测变量的中国宏观经济不确定性测度研究

本文提出利用大型统计数据直接测算中国宏观经济不确定性的方法。通过建立含有潜在不可观测变量的混频动态因子随机波动模型，实现利用月度和季度大型数据测度宏观经济不确定性。利用1994—2017年中国60个月度统计指标和4个季度统计指标，测算了我国宏观经济不确定性，结果表明:（1）中国宏观经济不确定性具有明显的阶段性特征，不确定性的变动受到多种因素的影响，传统的景气监测指标并不是一致最优的同步监测指标。（2）在测算宏观经济不确定性时，有必要将核心指标作为观测到的因子予以保留，不仅提高了结果的解释性，也能得到更符合经济事实的结果。（3）宏观政策变动是引起经济不确定性上升的重要因素，但政策的影响往往具有滞后效应，未预期的政策变动将触发更高的宏观经济不确定。

Measuring Macroeconomic Uncertainty of China Based on Large Dimension Data with Unobservable variables

This paper presents a method to directly measure the uncertainty of China's macroeconomic uncertainty using large dimension data. The macroeconomic uncertainty is measured by a stochastic volatility model of mixed dynamic factors with unobservable variables. China's macroeconomic uncertainty has been estimated using the 60-month and 4-quaterly statistical indicators from 1994 to 2017. The results show that: (1) China's macroeconomic uncertainty is affected by many factors and spikes around several years. The traditional indicators of macroeconomic climate index are not consistent of optimal synchronization indicators. (2) Retaining the core indicators as observed factors will improve the interpretability of the results and get more in line with the economic facts. (3) Macroeconomic policy is an important factor causing the economic uncertainty to rise. However, the influence of the policy often has a lagged effect. Unexpected policy changes will trigger higher macroeconomic uncertainty.