Hey guys!

Can anyone tell me where I can read more about the advances of Markov Switching in BVARS in Dynare? For example how I can use the commands in the Wiki for Markov Switching (e.g. dynare.org/DynareWiki/Markov … OptionsNew)? Or just point me to a help file, where I can read how to implement them, syntax, etc.?

I am new to dynare but had already a read through the manual and did some examples, so I got the basics of estimating DSGE models with Dynare (write “dynare model.mod” in the console and Pufff! - magic happens ). What I have now is a Markov Switching DSGE Model (yes, I know Dynare cannot handle that yet), which I have solved for the state representation:

X_{t} = G(St)*X_{t-1} + Pi(St).Z_{t}

Y_t=H_{t}*X_{t]

where St is just indicating the matrix has some Markov Switching coefficients. Now I was to the point, where I would like to estimate the Likelihood of that to do Bayesian (where I could write a Matlab version of Kim&Nelson’s 1999 code), but it relates to doing BVAR. (I guess Dynare uses K&N 1999?)

So far I have done this:

- I started here - dynare.org/DynareWiki/TableOfContents, looked through the MS part
- looked through the options - dynare.org/DynareWiki/Markov … OptionsNew
- currently working through the example (which says SWZ but is not SWZ, a similar model of a different economy/data - dynare.org/DynareWiki/example_swz

anywhere I can read more about using only parts of dynare in matlab file? (e.g. I solve my model in Matlab, get it in State space form and then ask dynare to only do the ML estimation, or do the it with Bayesian, but not through a mod file?)

Thanks!

P.s. Dynare version 4.2.4, Matlab version 7.12.0.635 (R2011a), Win 7, dunno if you need this.