I run johansen cointegration test on two time series variables using Eviews. the results reveal that there are two cointegration equations.
Is it right to interpret that there are two different patterns of trend between the two variables in the long run.
If you have two single time series variables as you have stated, then Engle & Granger is just the appropriate test for analysis. But when you analyze more than two variables Johansen cointegration test is recommended.
*
*
*
Good luck
If the results show two cointegrating vectors, the results can be questioned. The reasons is that the number of cointegrating vectors shall be one less than the number of variables. Since you have two variables, there should be one cointegrating vector with Johansen's test.
To correct this problem, select an appropriate lag order to run the test. Also, if there is a break in the time series, it should be considered in the analysis.
Two variable can have only one cointegrating vector not two. This means that the two variables may move away from the long-run average in the short-run, they will converge to long-run average in the longer period of time.
For more information, please check my papers available on the research gate to understand the cointegration test with and without a structural break.
I hope, it is helpful.
Good Luck!
Dr. Najibullah Hassanzoy
thank you Dr. Najibullah Hassanzoy for your answer
if there are four variables and two/three cointegration vectors, how to interpret the results.
in simple and non-technical language, how to interpret the two/three cointegration vectors between the variables.
You are welcome, my pleasure.
If you have three variable, two cointegrating vectors will be possible. This means that two long-run relationships are available among the variables. But, in practice, there may be one, two or no cointegrating relationship among three variables.
Each cointegrating relationship implies that there is a long-run relationship among the variables.
Good Luck!
Dr. Hassanzoy
i understood the meaning of two/three cointegration vectors. But, How to interpret the presence of two/three long-run relationships among the variabels. what exact does it indicate?
in other words, how to interpret two/three long-run relationships, so that a non-statistics person can also understand it.
thanks in advance
The long-run relationship means that the two variables will move together in the long-run while showing divergence in the short-run. In other words, the two variable are related to each other in someway. If you assume a line graph of the two variables, the lines will move around the long-run average closer to each other. This is the simplest meaning of cointegration.
I have explained the methods and interpreted the results in my papers, please download them from my research gate page for your more information.
Good Luck!
Hi Prof. Srikanth
You can use the following link which also uses E-views and the attached files to enhance your understanding of the Johansen cointegration test.
https://stats.stackexchange.com/questions/154256/johansen-cointegration-test-interpretation-of-results-in-eviews
Thank you and have a nice time
This video may be useful https://www.youtube.com/watch?v=O0auY_iDFIE.
You can use co-integration test in time series for long-run association or sustainability and average value such mean, median, standard deviation and variance change over the time.
That shows that the variables in question have long-run equilibrium relationship with each other,predict each other and likely granger caused each other
It is also used to find the number of relationships and also as a tool, to estimating those relationships in the variables to be estimated.
- Badges
- Science topic
- Similar topics
- Correction