A note on asymptotic properties of time series models with a trend break
A note on asymptotic properties of time series models with a trend break
复制标题
DOI：
10.2139/ssrn.3917796
复制
发表时间：
2021-09
期刊：
影响因子：
--
通讯作者：
Daisuke Yamazaki
中科院分区：
文献类型：
--
作者：
Daisuke Yamazaki
在本文中，我们重新分析Perron和朱（2005）的渐近性质的时间序列模型的趋势突变。我们证明了，对于具有平稳误差的联合突变趋势模型，当突变幅度固定时，他们的结果不成立。此外，我们表明，“收缩移位”的渐近框架是必要的，以建立这些结果。仿真结果表明，基于所提出的渐近理论的有限样本逼近效果良好。
In this paper, we re-analyze Perron and Zhu's (2005) asymptotic properties of time series models with a break in trend. We prove that, for the model with a joint broken trend with stationary errors, their results do not hold when the break magnitude is fixed. Furthermore, we show that the "shrinking shift'' asymptotic framework is necessary to establish these results. Simulation results illustrate that the finite sample approximation based on the proposed asymptotic theory works well.