两个平稳序列能组成一个cointegration系统嘛？

两个SETAR的gobal stationary的序列能组成一个cointegration系统嘛？或者进一步组一个threshold cointegration的系统？

虽然我知道书上说两个non－stationary的序列,可以组建一个cointegration，那么这两个non－stationary的序列有长期稳定关系。

那是不是说两个SETAR的gobal stationary的序列一定有长期稳定的关系（感觉不一定吧），如果不一定，在什么情况下它们有长期稳定的关系，这个长期稳定的关系用什么来表示。


另外，我们可以考察另外一种情况，比如一个三个regimes的SETAR（中间的regime是单根，根据书上的理论我们知道这个SETAR可以是平稳序列），例外一个序列是stationary AR， 我觉得这两个的差值能组成一个threshold cointegration的系统。换句话说，两个平稳序列能组成一个threshold cointegration的系统。大家觉得怎么样？

问得不错，望大家讨论，并提供文献和自己的理论分析。

xuehe 发表于 2012-4-2 20:29
问得不错，望大家讨论，并提供文献和自己的理论分析。

谢谢版大：）

能不能组成就需要检验，如果能从理论证明是最好的，
不能证明应该可以进行模拟，如果出现不稳定，或者类似于虚假回归，则两者就不能出现协整关系。

有谁能给出理论证明嘛？或在那本书，文章有

提供理论文献者奖励之！

From
Hansen, B., Threshold autoregression in economics. Statistics and Its Interface, 2011. 4(2): p. 123--128.

Balke and Fomby [ 7] introduced a “threshold cointe-gration” multivariate model which allows a discontinuous adjustment to long-run equilibrium. It is a combination of Tong’s TAR model and Engle and Granger’s model of cointegration [30] known as a vector error-correction model (VECM). The threshold cointegration model has been enor-mously influential in economics. Tsay [77] introduced a sim-ilar vector TAR (VTAR) model with possible threshold cointegration. Enders and Siklos [28] extended the Engle-Granger test for cointegration to allow for threshold adjust-ment. Estimation methods for the threshold VECM model were developed by Hansen and Seo [45], who also develop tests for the presence of threshold effects within the thresh-old VECM. Gonzalo and Pitarakis [ 33] develop similar tests in the context of a single cointegrating regression. Seo [ 71] develops a test for the null of no cointegration against the alternative of threshold cointegration. Seo [73] develops an asymptotic distribution theory for the estimates of the coin-tegration vector in the threshold VECM model.