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.
出处
《吉林大学社会科学学报》
CSSCI
北大核心
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)
关键词
GDP增速
经济预测
混频数据
动态因子模型
GDP grow th rate
economic forecast
mixed frequency data
dynamic factor model