主题Topic：A Simple Portmanteau Test with Improved FiniteSample Properties
时间Time：11月18号（周五）｜November18, 2016 (Friday), 14:00–15:35
地点Venue：文波207｜Room207，WENBO
主讲 / Speaker：
王学新博士，厦门大学王亚南经济研究院助理教授，2012年7月获西班牙马德里卡洛斯三世大学经济系博士学位，主要从事计量经济学、时间序列分析方面的研究。已发表文章在Journal of Time Series Analysis等国际知名学术期刊上。
Dr.Xuexin Wang, the Assistant Professor from Wang Yanan Institute for Studies inEconomics (WISE), Xiamen University, was graduated from Carlos Ⅲ University (Spain) , in2012. His academic papers are published in Journalof Time Series Analysis, etc.
研究领域 / Research Interests
计量经济学、时间序列分析
Econometrics，TimeSeries Analysis
摘要 / Abstract:
In this paper, we propose a simple Portmanteau test withbetter finite sample properties for general time series models with weakinnovations. This test has a quadratic form and follows a chi squareddistribution asymptotically. Its simplicity comes from an innovative transformof the sample autocorrelations of residuals, which allows for the weakinnovations and parameters estimation effect. Furthermore, a simplemodification is proposed to eliminate an important finite sample bias of the newtransform, improving the finite sample properties of the new Portmanteau test. Acomprehensive Monte Carlo simulation experiment demonstrates that the new Portmanteautest enjoys better size and power balance than other competing tests. Anapplication to the Standard & Poor 500 returns illustrates the merits ofour testing procedure.
