Changes to calculation of standard errors in --estat teffects-- after --sem--?

Hi all,

I am rerunning some data after about 6 months. I have been using sem and then estat teffects to determine the indirect effects of a variable through a mediator. I am unable to replicate my results from earlier. In particular, the coefficients and standard errors from the sem command are the same as before and the point estimate of the indirect effect is the same but the standard errors are different. Running the data in an older version (version 12.0) does not help. Does anyone know if the procedure to estimate indirect effects and their standard errors through ---estat teffects--- has changed? If so, what changed and why?

Thanks in advance!