Re: [Gretl-users] Panel data
On Tue, 24 Sep 2013, Hamza Bennani wrote:
Concerning panel data analysis, if I add data for multiple
individuals
in a stacked time series for a regression of the type :
Y(i,t)=A(i,t)X(i,t) + e. I obtain only one coefficient X at the end but
how can i manage to obtain a coefficient Xi for every individual (in a
random coefficient effect model).
Gretl doesn't offer built-in support for the random coefficients model.
But you could estimate this yourself using gretl's "mle" command to
estimate the covariance matrix.
Cameron and Trivedi (Microeconometrics, 2005, section 22.8.2) give an
exposition of the estimation procedure that would be quite helpful for
anyone wanting to program this in hansl (gretl's scripting language).
Allin Cottrell