MSVARSETUP / @msvarsetup compile error

[TomDoan]

No. There's no "Monte Carlo" in that for doing error bands. MCMC for Markov Switching models is very complicated because of all kinds of special issues with regimes getting too small, covariance matrices going singular, regimes switching places, etc. That's why the Ehrman-Ellison-Valla program for MCMC is over 200 lines while the two that just do point estimates are well under 100.

The Structural Breaks and Switching Models e-course goes into the technical details of MCMC for different types of MS models; in particular, it does a deep dive into the technical details of the EEV programs.

[shams11]

Thanks for all the help man...I hope I can eventually get it after looking at the e-course.