Formal MCMC Convergence Diagnostics such as Brooks-Gelman-Rubin Statistic after the bayes command in Stata 15

[John Ngombe]

Hello everyone! Is there a formal way (such as Brooks-Gelman-Rubin (BGR) statisticsor or Geweke Diagnostic statistic) in Stata 15 to determine convergence of Markov chain Monte Carlo (MCMC) if one estimates an econometric model using the bayes command in Stata 15? I am estimating a Bayesian Tobit Regression Model using bayes tobit: y, x1 x2, ul. However, Stata 15 seems to only rely on graphical methods to determine convergence of the MCMCs. I am interested in the formal tests but have not found anything yet on the same. Any help will be appreciated. Thank You!!

[Kjell Weyde]

Hi! I am also curious about tests of convergence after bayes, have you come up with any answers to this yet?

Kjell weyde

[Leonardo Guizzetti]

There is a user-created command, -grubin-, to compute the Brooks-Gelman-Rubin statistic. See https://blog.stata.com/2016/05/26/ge...ltiple-chains/

[bln tfte]

I did '' bayes, tratio: streg x1 x2, distribution(weibull) time ' and wanťo do model diagnostics using Gelman-Rubin method. Can anyone reach out?