我国经济增长速度的实时预测  

Real-Time Forecasting of Chinese Quarterly GDP Growth Rate

作　　者：陈磊[1] 隋占林 咸金坤 

机构地区：[1]东北财经大学经济学院

出　　处：《吉林大学社会科学学报》2016年第5期43-51,188,共9页Jilin University Journal Social Sciences Edition

基　　金：国家社会科学基金重大项目(15ZDA011);国家自然科学基金项目(71173029);辽宁特聘教授项目(2012)

摘　　要：通过构建混频动态因子模型对我国季度GDP增长率进行实时预测和评价,同时考察相关月度经济指标对预测GDP增速的贡献。研究结果显示:混频动态因子模型可以充分利用最新公布的相关月度和季度数据来提高GDP增速的预测精度;物价、进出口和工业生产等统计指标对改善预测精度的贡献较大,货币供应量等指标次之,而PMI等调查指标几乎没有贡献。当季第1个月公布的数据能较显著地改善GDP增速实时预测的精度,第2、3个月公布的数据的改进效果逐渐变弱。This paper real-timely forecasts Chinese quarterly GDP grow th rate by constructing mixed frequency dynamic factor model and investigates the contribution of different variables to forecasting GDP grow th rate. The research show s that the mixed frequency factor model could fully use the latest data w hich is gradually published to improve the precision of the forecast. Price,export and industry indicators get the lion's share of contributions,the variables of currency supply make less contributions and PM I make the least. The data published in the first month of a quarter could improve the precision of forecast greatly,and the improvement gradually w eakens w ith the data published in the second and third month.

关 键 词：GDP增速 经济预测 混频数据 动态因子模型 

分 类 号：F124.1[经济管理—世界经济]