Hi, my empirical strategy is as follows:
1. Apply coarsened exact matching to get weights
cem all_controls, treatment(cem_weights_from_step1)
2. Run xtreg with random effects but using mle option as it is the only that accepts weights estimated from step 1. As follows:
xtreg log_prices Covariates, re [iweight=cem_weights_from_step1]
Results are pretty cool but command from step 2 does not report/calculate R-squared. For comparability purposes with other estimations I need it. Is there any way to calculate it manually?
Best,
Mario
0 Response to Panel Random Effects/MLE, rho and R-squared
Post a Comment