Comparing estimated coefficients from different moments within panel data regression

Dear statalisters,

I hope you can help me out with this one. I have a panel data from 1998 to 2007. My equation to be estimated includes one dependent variable, four independent and two control variables. The dependent variable is dichotomous, so I estimated my model using logit regression:

logit RADIC DCT EPCT IILR IRI SIZE ABS

My question is the following: if I wanted to compare the results (i.e. coefficients) of two different years (let’s say 1999 and 2003), Could I do so within the panel data regression, that is, without ignoring the observations of the other years? Or is it better simply to estimate the same equation like a cross-sectional study in each of those years? In the latter case, I think that the commands would be:

logit RADIC DCT EPCT IILR IRI SIZE ABS if year == 1999
logit RADIC DCT EPCT IILR IRI SIZE ABS if year == 2003

But I want to know is how to do the first option, if it is possible and make sense (which I am not sure).

Thanks in advance

Best regards,
Antonio