摘要
揭示宏观经济与利率期限结构的混频作用机制有助于及时预判宏观经济形势,也是优化宏观政策跨周期和逆周期调控的基础和保障。本文融合中国独特的同比数据信息属性,提出一种包含日度信息的混频利率期限结构模型,即D-MF-NS模型。该模型引入累积器变量,避免可能出现的“维度灾”问题。结果表明,首先,与基准模型相比,即使包含新冠疫情期间的样本信息,日度混频模型对收益率曲线拟合效果依然占优。其次,通货膨胀率对高频利率期限结构水平因子有显著正向影响,其对水平因子预测方差在长短时间跨度中都有较大贡献,宏观经济跨周期和逆周期调节和操作均需要关注预期通胀机制。再次,日度高频斜率因子对GDP具有显著的正向影响,时间跨度无论是长期还是短期,收益率曲线斜率因子都是经济周期的稳定驱动因素。跨周期和逆周期政策的调控都需要重视收益率曲线斜率因子的作用机制。最后,我国货币政策在防通胀和促增长方面的跨周期调控效果明显好于逆周期调控效果。本文从数据频率的视角阐述利率期限结构的作用机制并为宏观经济政策的跨周期和逆周期调控提供新的经验证据。
Faced with the triple pressures of demand contraction,supply shock,and weakening expectations,the Chinese government has repeatedly raised the issue of macroeconomic cross-cyclical and countercyclical policies.Macroeconomic theory and practice teach us that the effect mechanism between China's macroeconomic and term structure of interest rates helps to meet the needs of cross-cyclical and countercyclical macroeconomic policy regulation.From the perspective of data frequency,the macroeconomic mechanisms of different cycles or time spans can be regarded as macroeconomic mechanisms of different data frequencies.Traditional macroeconomic theory and empirical research are often limited to quarterly or annual low-frequency mechanisms,making it difficult to apply a combination of cross cycle and countercyclical macroeconomic adjustment policies.In addition,the absolute time period of macroeconomic policy regulation within a single cycle may be relatively short,and macroeconomic policy evaluation faces the problem of small sample size.This paper advocates the construction of a mixed frequency data macroeconomic model to accurately identify the time-varying and smooth features of macroeconomic mechanisms under different time spans.The mixed model also has a unique advantage in overcoming the small sample problem.It can not only effectively utilize various available effective information,but also derive rich high-frequency information,thereby solving the problem of sample scarcity within a single cycle.This paper integrated the unique year-on-year data attributes of China and constructed a mixed frequency term structure model,namely the D-MF-NS model.The advantage of this model lies in the introduction of accumulator variables,which significantly reduces the data dimension of the mixed state space model,thereby avoiding potential“dimension disaster”problems.The results are as follows.First,compared with the benchmark model,our model has a better fitting effect on the yield curve even if it contains the sample informati
作者
尚玉皇
郑挺国
SHANG Yuhuang;ZHENG Tingguo(Southwestern University of Finance and Economics,611130;Xiamen University,361005)
出处
《财贸经济》
CSSCI
北大核心
2023年第11期88-105,共18页
Finance & Trade Economics
基金
教育部人文社会科学重点研究基地重大项目“数字金融与金融安全问题研究”(22JJD790069)
中央高校基本科研业务费项目“总体国家安全观视域下的金融风险防控理论机制与防范对策研究”(JBK230118)
国家自然科学基金面上项目“基于高维VAR的大型经济系统建模及应用研究”(72373125)。
关键词
利率期限结构
混频数据
跨周期-逆周期
通胀预期
货币政策
Term Strueture of Interest Rates
Mixed Frequency Data
Cross-Cyelical and Countereyelical
Inflation Expectation
Monelary Policy
