Panel data and Standard Error

Question

I currently have a set of panel data pertaining to a certain set of companies with a filters like minimum market cap, revenue size etc. If I expand this dataset by adding data about companies which don't necessarily are a part of the above mentioned filters, then will the standard error of a regression analysis increase?

Answer 1

This question is difficult to follow. What do you mean by "filters"? Are you referring to independent variables in a regression model? In any case, the standard errors should be cluster-robust standard errors, with the clusters being the individual companies. Adding more companies will add more clusters.

It would be good to have more info before going on, as it's not clear what you mean by "filters" and if this is a true panel data set with individual levels with time series components. You should also have a time-dummy to control for time fixed effects, along with using cluster robust standard errors. Wooldrdige provides several good discussions on panel data standard errors.

Answer 2

Yes it will increase because SE is inversely proportional to sample size. Adding more filters reduce the size and SE will increase.