Question about estimation: random number and seed

Hi all, I have a question regarding estimation. While reading forum, I noticed some said the estimation results changed using different computers and dynare versions. However, isn’t estimation based on MH algorithm drawing random samples? So even using the same machine and dynare version, there may be little differences between outputs. Isn’t that true?

No. Recent Dynare version set a seed for the pseudo-random number generator. That way, you can reproduce results.

1 Like