中国宏观经济情境设计与路径预测  

Scene Design and Path Forecast for Chinese Macroeconomic

作　　者：刘汉[1] 刘金全[1] 

机构地区：[1]吉林大学数量经济研究中心,吉林长春130012

出　　处：《中国管理科学》2013年第1期47-56,共10页Chinese Journal of Management Science

基　　金：国家社会科学基金重大项目(10ZD&006);国家自然科学基金项目(70971055)资助;教育部2011年"博士研究生学术新人奖";吉林大学2012年"全国优秀博士学位论文培育资助"项目资助

摘　　要：路径预测是通过构建联合置信区间来获得未来多期预测的可能范围,覆盖了中国宏观经济的可能运行路径。通过系统迭代和直接预测方法对我国宏观经济中的产出,通胀和利率进行蒙特卡洛模拟和路径预测,并在特定宏观经济情景下进行伪样本外条件路径预测。基于中国数据的蒙特卡洛模拟结果显示Scheffé方法构建的预测路径具有合理的路径覆盖范围,是判定预测不确定性的合理且有效方法,伪样本外条件路径预测显示合理的情境设计能够得出更为有效的预测路径。Path forecast refers to constructing simultaneous confidence regions of multi-step forecast, which is a reasonable summary of possible path in China economy. In this paper, iteration and direct forecast method is performed to construct the path forecast of output, inflation and interest rate and Monte Carlo simulation is performed. Pseudo out of sample path forecast with scene design of Chinese macroeconomic is also conducted. The results based on Chinese data show that, forecast path built by Scheff4 method has a reasonable coverage of the possible path, and it is a reasonable and effective way to forecast uncertainty of Chinese macroeconomic. The pseudo out of sample path forecast shows that a more reasonable and effective path forecast can be drawn with a reasonable scene design.

关 键 词：路径预测 不确定性 情境设计 VAR 局部投影 GDP 通货膨胀 

分 类 号：F224.0[经济管理—国民经济]