Hi there,

I am performing panel data regressions using the command xtreg including both time and firm-fixed effects based on robust standard errors. However, when doing so I get R-squared within of around 0.10, while the R-squared between is 1.000. The overall R-squared is therefore also very high, i.e. 0.78. When performing the same regressions with fixed effects using either reg or areg, I get the same high R-squared value of 0.78. When leaving out the firm-fixed effects though, the R-squared between takes on a value of about 0.40, while the overall R-squared value using both xtreg and the other commands also becomes much less extreme. Is the value of R-squared between of 1.000 an indication that I should not use firm-fixed effects or how should I interpret these values.

Thank you for your response!

Kind regards,

Sara