say you have 20 observations across time on 10 individuals each.
Assuming that the unobserved effects are uncorrelated with the
independent variables, what model specification would be
appropriate for the following estimation:
(i) individual specific intercepts differ, slopes across
individuals are equal but the variance of the error term differs
across individuals
(ii) individual specific intercepts differ, slopes across
individuals are equal but the errors are contemporaneously
correlated across individuals in the same time period