Functional-coefficient cointegration models in the presence of deterministic trends
Functional-coefficient cointegration models in the presence of deterministic trends
复制标题
DOI：
10.1080/07474938.2015.1092845
复制
发表时间：
2018-05
影响因子：
1.2
通讯作者：
Masayuki Hirukawa;Mari Sakudo
中科院分区：
文献类型：
--
作者：
Masayuki Hirukawa;Mari Sakudo
本文将函数系数协整模型（FCCM）推广到非平稳回归变量同时包含随机和确定性趋势的情形。探讨了FCCM局部线性（LL）回归平滑器的非退化分布理论。它表明，即使当集成回归是内源性的，极限分布是相同的，如果他们是外源的。通过Monte Carlo模拟与替代估计方法相比，LL估计的小样本性能进行了研究。作为FCCM的应用，考虑了伊利诺伊州的电力需求分析。
ABSTRACT In this article, we extend the functional-coefficient cointegration model (FCCM) to the cases in which nonstationary regressors contain both stochastic and deterministic trends. A nondegenerate distributional theory on the local linear (LL) regression smoother of the FCCM is explored. It is demonstrated that even when integrated regressors are endogenous, the limiting distribution is the same as if they were exogenous. Finite-sample performance of the LL estimator is investigated via Monte Carlo simulations in comparison with an alternative estimation method. As an application of the FCCM, electricity demand analysis in Illinois is considered.