基于状态空间模型的中国季度GDP季节调整(1996～2009年)  

The Quarter Seasonally Adjusted GDP in China Based on State Space Model

作　　者：桂文林[1,2] 韩兆洲[1] 

机构地区：[1]暨南大学统计系 [2]惠州学院数学系

出　　处：《数量经济技术经济研究》2011年第7期77-89,共13页Journal of Quantitative & Technological Economics

基　　金：国家哲学社会科学基金重点项目(07AJL009);国家哲学社会科学基金(10BJY050);广东省哲学社会科学基金(09E-04)的资助

摘　　要：中国迄今为止尚未公布包括季度GDP在内的经季节调整的经济指标数据,这不仅不利于对中国宏观经济运行监测,也无法满足国际比较的迫切需要。本文对中国1996年一季度至2009年四季度的实际GDP构建基于状态空间形式的季节调整模型,通过卡尔曼滤波递推算法对状态向量的各分量进行了最优估计、平滑和预测,并对超参数进行了极大似然估计。在此基础上分析了这一期间中国GDP的主要季节和趋势特征,并计算出了季节调整后的季度环比增长率指标用来分析和监测经济走势,鉴别趋势拐点,制定相关经济政策。最后通过与国际通用的TRAMO-SEATS季节调整模型的对比发现其优越性。The seasonally adjusted economic indicators including quarterly GDP have so far not been announced in China. This is not only conducive to the monito- ring of China＇s macro - economy, but also can not meet the urgent need for interna- tional comparison. In this paper, after building a seasonal adjustment model in the form of state space for China＇s real GDP from 1st quarter 1996 to 4th quarter 2009, getting the optimal estimation, smoothing and predicting the various components of state vector through the Kalman filter recursive algorithm, the super - parameters have been estimated by maximum likelihood method. On this basis, the main characteristics of the trends and season of China＇s GDP during the period are analyzed. The quarterly sequential growth rate of seasonal adjustment is calculated to analyze and monitor economic trends, identify trends inflection point and the development of relevant economic policy. Finally, by the contrast to international generic TRAM O-SEATS model for seasonal adjustment, its advantages are found.

关 键 词：状态空间模型 季度GDP 季节调整 卡尔曼滤波 

分 类 号：F061.2[经济管理—政治经济学]