hi everyone, I met one problem when I use the SFA model to calculate the TFP. I am very grateful if you can provide some advice.
my code:

global xvar_prod lnx1 lnx2 lnx3 lnx4 t lnx1x1 lnx1x2 lnx1x3 lnx1x4 lnx1t lnx2x2 lnx2x3 lnx2x4 lnx2t lnx3x3 lnx3x4 lnx3t lnx4x4 lnx4t tt
sfmodel lny,prod dist(h) frontier($xvar_prod) usigmas(t) vsigmas()
ml max,difficult nrtol(1e-5) gtol(1e-5)
initial: log likelihood = -72595.097
alternative: log likelihood = -42600.451
rescale: log likelihood = -1524.7414
rescale eq: log likelihood = -843.91446
varlist required
r(100);

Thanks.