A SAS Macro for Estimation of Unbalanced Panel Data Models with Two Fixed Effects

[#1]

This work aimed at presenting a faster general algorithm for estimation
of panel data models with two fixed effects, when it is compared with PROC
PANEL, working with balanced or unbalanced data. The results by simulation
showed that until with a large data base (50,000 observations) only a few
seconds were necessary for estimation, while in the PROC PANEL, the estimation
wasn’t possible due to out of memory. You can find more details on paper SGF 354-2011

http://support.sas.com/resources/papers/proceedings11/354-2011.pdf

[Joakim]

It's great that you have adressed this problem with unbalanced data set. But this macro only estimate a Two-way fixed effects. If I want to perform a one-way random effects estimation with an unbalanced panel I still run into the problem of trying to estimate a too large set of individual dummiy variables.

Do you or anyone else know how to perform a one-way random effects with N>50000 and T between 2 and 7 (i.e. a large unbalanced panel)?

Kind regards

Joakim