Marginal data density computation

Hi all,

I have a quick question. While computing the marginal data density of a model after estimation, dynare uses the MHM method of Geweke. Geweke proposed a truncated multivariate normal distribution as the density to use for the integration.

I was wondering how Dynare decides where to truncate the distribution?

Thanks. I appreciate your help.

Hi Openmacro,

How do you know about the MHM method of Geweke and that he proposed a truncated multivariate normal distribution. Could you provide me link if there is a particular website where you got this information from. I would like to read more on this topic and then I can go to investigate the Dynare source code in order to answer your question.

Thanks in advance!