Panel Data Regression with Groups

[Hakan Self]

Hello,
i would like to make a panel data regression with -reghdfe- considering country specific fixed effects. I have 20 countries over 5 years. I want to group the countries, some being advanced countries and some being emerging countries.
what would my analyses look like and would be there a benefit for grouping them ?

xtset country_id Year

reghdfe y x1 x2, absorb(country_id) vce(cluster country_id)

[Jared Greathouse]

No idea what you mean here- group them how from an analytic standpoint?

[Carlo Lazzaro]

Hakan:
I do share Jared's concerns about the robustness of your approach.
That said, you can simply create a new -panelid- (ie, a categorical variable that groups countries togeteher according to a given set of criteria), -xtset- your dataset that way and replace -country_id- with the new -panelid- when necessary in your regerssion equation.
It is also important to stress that your -panleid- changes substantively upin following the approach you ahe in mind: from countries to group of countries sharing a given set of features.