Year gaps in unbalanced panel data set

[Constantin Domizlaff]

Hey all,

I am working with an unbalanced panel data set with total corporate carbon emissions as the dependent variable and a few board characteristics as predictor variables. Due to the low data availability of the emissions figures, the panel is unbalanced. I already dropped observations that would lead to a difference in sample size between my different models as I am also running alternative regressions with different dependent variables.
Now, when I use the -xtset- command to set my panel data it returns that my time variable (year) includes gaps. Are those gaps potentially problematic, and if so, how can I remove those gaps or handle them in another way? I know there are only few of those gaps, but I wanted to ask nonetheless.

Thanks a lot in advance!

[Nick Cox]

It's the other way round. The xtset command helps you by ensuring that calculations, include model fits, do sensible things given the gaps. The only ways to remove gaps that I know of are to renumber dates ignoring them (quite wrong), to omit incomplete panels (often a poor strategy), or to fill gaps by interpolation or imputation (never totally unproblematic).

[Constantin Domizlaff]

Hi Nick, thanks a lot for the quick answer. I do not want to mess with the dataset too much or overmanipulate it so what I read from your response given what I wrote I can work with the gaps and run my models normally?

[Carlo Lazzaro]

Constantin:
yes, you can.

[Nick Cox]

In short, yes indeed.