Bayesian MCMC Gibbs sampling

[Jie Xing]

I want to use Bayesian MCMC to generate a sequence of draws from the posterior distribution. I am using Stata 14.1. I know the command bayesmh which uses the MH,

• what is the command using Gibbs sampling to simulate draws sequentially for blocks of parameters?
• How to incorporate instruments within Bayesian framework?
• Could you give me a simple example and corresponding codes?
• Or are there any papers to refer?

Appreciate it so much!

[Matthew J. Baker]

Jie Xing --

What sort of problem do you have in mind? Without more information about the setting, it is hard to suggest anything further. If you are writing your own likelihood in Mata, you might take a look at my paper.

Best,

Matthew J. Baker

[Jie Xing]

Matthew J. Baker
Hi Matthew J. Baker,

Thank you so much for your kind advice and sharing the link of your paper.

The problem is to approximate the distribution of parameters and some latent variables for three equations which respectively describe demand, pricing and export market participation. To control for the endogeneity of the output price, I want to incorporate instruments within Bayesian framework. MJ Roberts et al.(2017) said Bayesian MCMC MH and Gibbs sampling can together achieve the results. But I don't know how to run in Stata 14.

I am reading your paper now, but I'm not sure whether your algorithms can resolve this problem.

Could you give me some more advice? Many thanks.

Best regards,
Jie Xing