HP filter (pooled vs fixed effect model)
I am using panel data over 290 Swedish municipalities, from year 2003-2013. My dependent variable is out-migration and my two independent variables are vacancy rate and unemployment rate. I am planning on running models using first pooled and then a fixed effects model, where the pooled will act as a baseline model.
I have used the HP filter to clear any trends, since my tutor has advised me so.
However, when I run the pooled regression and then the fixed effects regression with my filtered variables (xtreg out_migration vacancy_rate unemployment_rate, fe) I get the exact same numbers as the pooled regression model.
If I do the same thing with the un-filtered variables, I get different results.
Why do I get the same results?
I have used the HP filter to clear any trends, since my tutor has advised me so.
However, when I run the pooled regression and then the fixed effects regression with my filtered variables (xtreg out_migration vacancy_rate unemployment_rate, fe) I get the exact same numbers as the pooled regression model.
If I do the same thing with the un-filtered variables, I get different results.
Why do I get the same results?
