Hello,

We have the following dataset: A limited panel in which an establishment is found in the sample for only 5 years but we have a pooled cross-sectional time series spanning a very long period of time, from 1990 to 2015.

On this dataset we analyzed xtreg regressions with a fixed effect for establishment. The dependent variable is the mean value of wages within establishment-years. The baseline model only has controls for years.

Any help with interpretations of the results will be welcome. In particular am I right that:
1) The establishment fixed effect accounts for changes in the mean of wages over the “life-span” of the establishment within the sample.
2) The controls for years then capture the change in the mean value of wages between establishments over time.
3) Sigma_u captures the mean value of wages between establishments over time.
4) To measure between-establishments wage inequality in the year 2000, for example, I simply need to add the coefficient for the year 2000 to the value of sigma_u.