Page 1 of 1

Estimating BSP Model of Rabanal-Rameraz 2005

PostPosted: Wed Apr 03, 2013 9:44 am
by wmmahrous
Dear all,
Here I'm attaching Dynare code for estimating Rabanal-Rameraz BSP model (2005). I'm trying to modify the model by getting rid of real wages as an observed variable, because I don't have enough data about it for the country I am gonna apply this model on (Tanzania). Is it right to skip any shock (because the model has 4 shocks and four observed variables) to balance for the new number of observed variables? or Do I have to skip a specific shock? In my attached file you will find that I'm skipping the mark-up price shock...I'm waiting for your replies.
Thanks in advance,
Walaa Mahrous.
BSP model.mod
(1.07 KiB) Downloaded 240 times

Re: Estimating BSP Model of Rabanal-Rameraz 2005

PostPosted: Fri Apr 05, 2013 6:15 pm
by jpfeifer
In general, to avoid stochastic singularity, you need at least as many shocks as observed variables. Having more shocks should be no problem. So you could just drop the variable from var_obs and leave everything else unaffected. If you want to omit a shock, use one that is empirically not important.

Re: Estimating BSP Model of Rabanal-Rameraz 2005

PostPosted: Fri Apr 05, 2013 7:51 pm
by wmmahrous
jpfeifer....Thank you so much...I really appreciate your advice :)