I am in the process of running two different xtabond2 models, and have a question regarding the specification of “small T large N”. Our first model has 1880 employees over 9 years (T=9, N =1880, or around 4800 observations, total). Our second model has 1880 employees over 342 time periods (so T=342, N=1880, or around 100,000 observations, total).

I am wondering if there is a rule of thumb regarding what is meant by “small T large N”. If xtabond2 is not appropriate for the second model, can you recommend an alternate specification?

These employees belong to 32 firms and we intend to include firm fixed effects, in addition to time fixed effects in both models. Is the inclusion of firm fixed effects acceptable in the xtabond2 model?

Thanks for your help.