Comparing coefficients in separate regressions from different time periods

[Giorgio Depolli]

Dear all,

I want to compare the coefficients of the variable x from 2 different time periods and see if they are statistically different from each other. I use the code:

.reg emp x y z i.year if year>2003

.estimate store m1

.reg emp x y z i.year if year<2004

.estimate store m2

.suest m1 m2, cluster(region)

But Stata gives me an error saying that " year: factor variable base conflict". Is this because regressions are run on different time periods? is there a way to fix it and then test if coefficients of x are the same?

Could you suggest any alternative ways?

[Andrew Musau]

Try:

reg emp x y z i.year if year>2003
estimate store m1
gen yr= year
reg emp x y z i.yr if year<2004
estimate store m2
suest m1 m2, cluster(region)

[Giorgio Depolli]

Thank you Andrew for your help. It works perfectly for OLS.

However, do you know how to do suest with -ivreg-?

e.g.

ivreg2 emp (x=x1) y z i.year year>2003

....

[Andrew Musau]

suest does not work with ivreg2 (from SSC), but the following may help.

https://www.statalist.org/forums/for...ferent-samples

[Giorgio Depolli]

Thank you Andrew it worked well!