First Differencing with Dummy variables

[Marcus Arkwright]

I have set my panel data up using: xtset personid date_monthly

I have run a regression with this data using many dummy variables for the month, education, marriage, health insurance etc. so that it looks like this.

reg lhours lwage i.month refyear edu_lessthanhs edu_hs edu_somecollege edu_college edu_more married healthins erp if female == 1.
I want to first difference this equation. When I use this command, only ref year and edu_more are dropped, but I would have thought more would be.
reg D.(lhours lwage month Quarter refyear edu_lessthanhs edu_hs edu_somecollege edu_college edu_more married healthins erp) if female == 1, vce(robust)

Is there a better way for me to do this. The variable date_monthly covers 48 months, but the panel isn't balanced.

Thank you,

Marcus Arkwright

[Maarten Buis]

If you want to estimate a fixed effects regression, then you gen just type xtreg y x1 x2 ..., fe No need to difference yourself.

[Marcus Arkwright]

Hi Maarten, I have used FE as well, but would like to see which method is more efficient, hence my need to first difference. In my differenced equation I have removed the i. in front of month and quarter to have it as a continuous time. Does that work in the context of this regression