基于多源高维数据的经济景气实时监测研究
Real-time Monitoring of Business Tendency Based on High-dimensional Data from Multiple Sources
唐晓彬 1崔茂生 1刘思明1
作者信息
摘要
为提升我国经济景气监测的时效性和准确性,不同于已有的以工业增加值同比增速为基准循环指标的景气指数体系,本文提出以GDP同比增速为基准循环指标的经济先行指数构建方法.首先构建混频Granger因果检验从多源高维数据中筛选出对于GDP同比增速具有稳定领先性的先行指标,在此基础上根据我国宏观经济数据特征,设计了一种面向高维缺失数据集的同比形式混频动态单因子模型来合成我国经济先行指数,最后对该指数的性能进行验证并应用该指数对我国经济景气状况进行实时监测和预测分析.研究结果表明,与现有的先行指数相比,本文构建的经济合成先行指数能够更加全面准确地反映我国经济运行状况,且在多数时期对GDP同比增速有稳定的领先性,提升了GDP同比增速预测的精度.经济合成先行指数的运行趋势在不同时期呈现出显著的差异,反映出近年来我国经济增速换挡、经济周期平稳化的阶段性特征.经济合成先行指数能够基于实时发布的数据流进行更新与调整,及时有效地预判经济走势的变化,具有较强的时效性.本文丰富了宏观经济监测相关研究,对构建高质量宏观实时监测系统具有借鉴意义.
Abstract
In order to enhance the timeliness and accuracy of economic prosperity monitoring in China,different from the existing economic prospering index system that uses year-on-year industrial added value growth rate as the benchmark target,this article proposes a method for constructing a leading index based on year-on-year GDP growth rate as the benchmark cyclicat indicator.Firstly,the Mixed-frequency Granger causality test is applied to select stable leading indicators from a multitude of high-dimensional dataset.Secondly,a Mixed-frequency Dynamic Factor model on high-dimensional macroeconomic dataset with arbitrary pattern of missing data is designed to construct the leading composite economic index,taking into account the characteristics of macroeconomic data.Finally,the performance of the economic composite leading index is validated,and it is applied to real-time monitoring and predictive analysis of China's economic conditions.The results show that,compared with the existing leading index,the constructed economic composite leading index with year-on-year GDP growth rate as the benchmark target can accurately reflect the overall economic conditions and improves the accuracy of year-on-year GDP growth rate forecasting.The economic composite leading index varies significantly in different periods,reflecting the phased characteristics of China's economic growth shift and the stabilization of economic cycles since entering the new normal.The economic composite leading index can be updated and adjusted based on real-time data streams,enabling timely and accurately economic forecast.This study enriches the research related to economic monitoring and provides valuable insights for constructing a high-quality real-time economic monitoring system.
关键词
经济景气/混频Granger因果检验/混频动态因子模型/实时监测Key words
Business Tendency/Mixed-frequency Granger Causality Test/Mixed-frequency Dynamic Factor Model/Real-time Monitor引用本文复制引用
出版年
2024
NETL
NSTL
万方数据