摘要
对于一个实际的时间序列数据,我们并不知道其真正的数据生成过程,只能通过假设和基于假设的统计推断来确定。本文利用蒙特卡罗模拟进行有限样本的ADF检验。结果表明,当原假设是存在单位根,备择假设是趋势稳定时,ADF检验的势很低。这意味着,我国GNP存在单位根是不确定的。
We have to draw statistical inference by hypothesis testing, because we don't know the real data generation process of any time series. The author applies to ADF test on monte carlo simulation. The result shows that the power of ADF test is low if null hypothesis is unit root and the alternative is trend stationary. It suggests the uncertainty about the existence of a unit root in real output and the uncertainty about the amount of persistence of macroeconomic shocks.
作者
汪燕敏
柯健
WANG Yan-min, KE Jian (Economic Development Research Center, Anhui University of Finance and Economics, Bengbu 233041,China;School of Information Engineering, Anhui University of Finance and Economics, Bengbu 233041,China)
出处
《电脑知识与技术》
2008年第5期723-725,736,共4页
Computer Knowledge and Technology
基金
安徽省高校自然科学资金资助(KJ20088138),安徽省高校青年教师社科项目科研资助(2006jqw075)
关键词
数据生成过程
单位根检验
蒙特卡罗模拟
data generation process
unit-root testing
monte carlo simulation