Dear All,

I have micro data at firm level.
I am looking to the impact of migration on firm-level productivity. I am just wondering what's the intuition behind modeling and trying to quantify the impact based on Case 1 and case 2.
When should I add treatment as a seperate regressor?

Many thanks in advance,
Nazlı

***Post Treatment Period:
gen Time=1 if year>2002
replace Time=0 if year<2003

**Interaction term:
gen TT=Time*Treatment

Case 1:
reg lnlp Treatment TT i.year
reg lnlp TT i.year

Case 2:
reg lnlp Treatment TT i.year labourforecgrowth(city level)
reg lnlp TT i.year labourforecgrowth(city level)