信心、预期和中国经济周期波动

Confidence, Expectation, and China’s Business Cycles

有关文献大多将预期视为不可观测的潜变量,本文尝试考察预期的可观测性。首先采用混频时变DFM方法估计了中国的潜在产出,以代表经济基本面,发现中国经济在2008年金融危机前就开始下行。接着构建了一个不完全信息SVAR模型,基于动态识别方法估计预期冲击和噪声冲击的影响。研究发现:消费者信心指数“错误”预期主要宏观变量的变化;相较而言,商业信心指数是更好的预期指标,但包含噪声成分。以商业信心指数为预期变量,预期冲击对主要宏观变量形成持久的正向效应,预期冲击解释了潜在产出变化的60%,解释了GDP变化的55%;噪声冲击对宏观变量具有正向“驼峰状”影响,对GDP的短期解释力超过8%。消费者信心指数表现不佳可能源于消费和宏观经济的脱节。本文的基本结论在大数据集、替代指标、基本面检验以及其他识别方式下依然成立。

Most existing news-shock research literature regards expectation as an unobservable latent variable,but in this paper,we try to examine its observability.We estimate China s potential output by mixing-frequency DFM method to represent the economic fundamentals.We find that China s economy began to decline before the 2008 financial crisis.Then we employ an imperfect information SVAR model to estimate the impact of news shocks and noise shocks based on a novel dynamic identification method.We find that the consumer confidence index contains the incorrect expectations of the major macro variables;in contrast,the business confidence indicator is a better one but contains many noise components.The news shocks account for nearly 60%of potential output variation,and 55%of GDP variation.The noise shock has a positive effect on macroeconomic variables,forming a“hump”effect,which has a short-term explanatory power for more than 8%of GDP.The poor performance of the consumer confidence index may be due to the disconnection between consumption and macro-economy.The conclusions still hold under large data sets,alternative indicators,fundamental test and other identification restrictions.