Country specific time trends with first-differenced regression

[Chiara Piazzo]

Dear all,

I am running a first-differences regression and was wondering if country specific time trends can be used as country fixed-effects should not be used in a FD regression?

I have data from 1985 to 2010 so I compute generate time trends as follows;

gen time = year - 1985
gen time2 = time^2
qui xi i.iso3_code*time, pref(_yi_)
qui xi i.iso3_code*time2, pref(_y2_)

I would like to run the following regression;

xi: reg Y_yi_* _y2* X1 X2

Is this an appropriate way to proceed? I realize that this is not a strictly Stata question but any help will be greatly appreciated. Thank you!

Sincerely,

Milu

[Phil Bromiley]

With panel data, you're almost always better off using panel estimators. If you used xtreg, you could include c.year#i.country in the regressors.

[Chiara Piazzo]

Thank you, Phil Bromiley! Appreciate it.

Sincerely,

Milu