如果r = 4,需要多少限制来测试协整关系?
我估计了一个协调一致的五个I(1)变量的VECM。该测试导致四个协整方程。现在,我想确定所有协整的向量,我的问题是:我最少需要进行多少限制,以及如何使用VECM()函数执行LR检测?
VECM1 = VECM(vardata1,lag=0, r=4, estim = c("ML"),exogen = M,include = "none")
H=matrix(byrow=T,c(1,0,0,0,0,
1,1,0,0,0,
0,0,1,0,0,
0,0,1,1,0,
0,0,1,0,1))
可悲的是blrtest(vecm1,h = h,r = 3)
仅适用于ca.jo对象。但是,使用ca.jo()
只能以最小var滞后顺序为2(我有1)估算VECM。
I estimated a VECM with five I(1)-Variables that are cointegrated. The test results in four cointegrating equations. Now I want to identify all cointegrating vectors and my question is: How many restrictions do I need to impose at minimum and how can I execute an LR-test using the VECM() function?
VECM1 = VECM(vardata1,lag=0, r=4, estim = c("ML"),exogen = M,include = "none")
H=matrix(byrow=T,c(1,0,0,0,0,
1,1,0,0,0,
0,0,1,0,0,
0,0,1,1,0,
0,0,1,0,1))
Sadly blrtest(VECM1,H=H,r=3)
only works for ca.jo objects. But with ca.jo()
a VECM can only be estimated with a minimum VAR lag order of 2 (I have 1).
如果你对这篇内容有疑问,欢迎到本站社区发帖提问 参与讨论,获取更多帮助,或者扫码二维码加入 Web 技术交流群。

绑定邮箱获取回复消息
由于您还没有绑定你的真实邮箱,如果其他用户或者作者回复了您的评论,将不能在第一时间通知您!
发布评论