How to detect sources of variations in coefficients obtained from a panel data regression?

[Cooper Felix]

Is it possible to identify whether and how cross-sectional variation and/or time-series variation contributes to the coefficients estimated from a panel data regression (e.g., xtreg, fe or xtreg, re)? If so, which command can be used to detect this? Thanks!

[Clyde Schechter]

You can use -xthybrid-, Francisco Perales and Reinhard Schunck, available from SSC. This will explicitly separate within- and between- variation.