宏观数据发布与经济周期实时测度方法研究

发布时间：2018-06-26 12:12

本文选题：经济周期 + 宏观数据发布　；参考：《系统工程理论与实践》2017年04期

【摘要】：为充分利用实时发布的最新数据来改善宏观经济分析的时效性,本文扩展了一种能够处理不规则数据的混频区制转移动态因子模型及其贝叶斯估计方法.数值模拟结果表明贝叶斯方法提高了模型估计的准确性,并发现含噪音成分低的指标,其新数据对实时测度的贡献更大.基于2008年以来的256组实时数据的研究结果表明,文中模型不仅较好刻画了1992年以来我国经济周期波动及阶段性变化,而且对GDP数据修订具有很好的稳健性,,此外对经济周期拐点的实时识别则存在2至8个月的滞后.最后,在每月依序发布的指标中,进出口数据含有较高噪音成分,工业增加值和财政税收等数据对当月经济状况测度的更新修正幅度大且可靠性高,对于提高经济周期测度时效性具有重要价值.
[Abstract]:In order to make full use of the latest data released in real time to improve the timeliness of macroeconomic analysis, this paper extends a mixed region dynamic factor model and its Bayesian estimation method, which can deal with irregular data. Numerical simulation results show that the Bayesian method improves the accuracy of the model estimation and finds the index with low noise content. The new data has a greater contribution to real-time measurement. Based on 256 sets of real time data since 2008, the results show that the model not only describes the economic cycle fluctuations and periodic changes since 1992. Moreover, the revision of GDP data has good robustness, and the real time identification of the inflection point of the economic cycle has a lag of 2 to 8 months. Finally, in the index published in order of each month, the import and export data contain relatively high noise content, industrial added value and fiscal and taxation data have a large correction range and high reliability to the economic condition measure of that month. It is of great value to improve the timeliness of business cycle measurement.
【作者单位】：厦门大学经济学院统计系;厦门大学王亚南经济研究院;
【基金】：国家自然科学基金(71371160) 教育部新世纪优秀人才支持计划(NCET-13-0509);教育部“长江学者奖励计划”青年学者项目~~
【分类号】：F124;F224

【相似文献】