Bayesian Estimation with change in policy Rule

[cmontoro]

Dear all:
I need to estimate a model with a change in the policy rule, that is:
RR=(1-I_IT)*RR1+I_IT*RR2;
RR1=PHI_P1*DP(+1)+PHI_Y1*GAP+MON1;
RR2=PHI_P2*DP(+1)+PHI_Y2*GAP+MON2;
where I_IT is a {0,1} indicator which is exogenously determined (that is change in the monetary policy regime).

How should I declare the I_IT variable? and how should I give the data for this variable for the estimation? I've seen there is an option for varexo_det variables for forecast, but I´m not sure this works also for estimation.

Thanks,

Carlos

[MichelJuillard]

No such device is currently available in Dynare. I had several requests for it and we should implement it relatively soon.

Note however an important distinction. In forecasting with varexo_det, agents know the future value of the exogenous variables.

In an estimating framework, we want often to estimate under the hypothesis that the change of regime wasn't expected. This is what I'm planing to implement first.

Estimating under the hypothesis that the change of regime is anticipated is more complicated because it requires to provide Dynare with data about future expectation of regime change for each date in the sample.

Best

Michel