Pooled and time series data in one model
Posted: Sun Jul 13, 2014 3:01 pm
by Dthurley
I am running a model which has a mix of pooled data (dependent and one independent variables) and time series (two independent variables). I would like to use pooled data analysis instead of running a time series on each cross sectional unit/country. Is that allowed?
Re: Pooled and time series data in one model
Posted: Wed Jul 16, 2014 5:27 am
by nishantvats12
It depends on the length of the time taken under consideration. If it is like for 10 years or more it is advisable to go for time-panel econometric techniques.
Also, when going for a simple panel model, you may not restrict yourself to pooled model alone, you might want to go for Fixed and/or Random Effect Models to get a better outcome. The thought of which method to use depends on the objective(s) of a researcher and/or statistical methods like the restrictive f test and the Haussman test.