I am a new user. I was wondering whether dynare can deal with permanent (technology) shocks in a stochastic model. Any help will be much appreciated!

Yes, it can. See for example the attached Dynare example (taken from the examples and changed to run in the most current Dynare version), where y_obs and p_obs are non-stationary.

fsdat.m (4.52 KB)

fs2000ns_steadystate.m (1004 Bytes)

fs2000ns.mod (2.76 KB)

Thatâ€™s awesome, many many thanks!

i still cant find the permanent shock in his attach, isnt he giving a estimation but not a solution ???

If you have a process with a unit root like

`x=x(-1)+epsilon;`

where epsilon is a stochastic shock process, any shock epsilon will be permanent. However, the most common way to implement this into Dynare is to stationarize the model by detrending it, i.e. differencing the system. In this case, you will end up with

`dx=epsilon;`

This is similar to the dA in the attached mod-file. However, you might try to directly use a non-stationary shock process like the first one.