I have a time series data from 1988 to 2019 for GDP, Gross Fixed Capital Formation and Inflation rate, the variables were I(1) and I(0).
I used Engle-Granger to test for the long run relationship among them, but the results suggested that, no long run relationship exits among the variables.
Can I estimate the model using OLS multiple regression, or should I estimate the VAR model?
And what does the VAR do!
Thank you very much @Sinan
What is the test command for ARDL bound test in STATA or Eviews.
And honestly, these two models (VECM and VAR) are really confusing me.
May you please assist on when they are used and why!
Suleyman Abas
A VAR in levels can be estimated if you don´t find any cointegration between the endogenous variables in your system. If you find that they are cointegrated, you estimate a VECM.
To use the ARDL bound test, and you estimate a model with two variables, one of the variables must be weakly exogenous.
Thank you very much @Sinan your answers were really helpful.
And thank you very much all, I really appreciate your answers, be blessed
one of the solutions is a difference process to make data have the same order. If variables are cointegrated you can choose VECM model or using ARDL or VAR in level if data have no cointegration
- Badges
- Science topic
- Similar topics
- Linguistics
- Composition Studies
- Publication