Page 1 of 1

stationarity test

Posted: Sun Dec 13, 2020 1:24 pm
by lisa
Hi,
I will estimate a VAR model with 3 variables in logarithms (log(gdp), log(inflation rate) , log(interest rate)).
My question is do I have to test the stationarity of variables (gdp , inflation rate and interest rate) or the stationarity of the logarithms of variables (log(gdp) , log(inflation rate) and log(interest rate).
Thank you for your help.

Re: stationarity test

Posted: Thu Dec 17, 2020 11:54 am
by JuanNicolasDamico
Hi Lisa, you have to perform the test in the log transformation.
There is a paper for example written by ender and Lee that I can recall now where they do unit root testing on the variables in logs.
If you need help with performing unit root tests you can check my video: https://www.youtube.com/watch?v=65g6D4bICQY
Feel free to ask further questions, Also: here is a step by step how to estimate VAR models in Eviews and do variance decompositions, IRF, granger causality test, etc. https://youtu.be/dWiITAmSP5k
Regards,

Re: stationarity test

Posted: Fri Feb 19, 2021 8:35 am
by lisa
Thank you for your reply.
But the last video is not available.