Transitory shocks in a stochastic model

Hi Michel and people checking out the forum,

I am working with an extended version of Smets and Wouters (2007).

What I want to get is the impulse response of consumption to a government spending shock, when the nominal interest rate is at zero for the first 10 periods.

To do this, I created a deterministic exogenous variable “indicator” which is going to multiply my monetary policy equation, and I set this variable to 0 for the first 10 periods, and then back to its initial steady state value of 1:

var indicator;
periods 0:10;
values 0;

However, when I simulate the economy using simul_stoch, the interest rate does not get set to zero for the first 10 periods, for some reason. The simulated model just gives me a zero interest rate for the first period, but it does not seem to obey the above constraint I set on “indicator”.

I tired using the forecast function, but the results are worse. I get graphs which show no movement at all in the endogenous variables. I don’t really know what is going on with that.

I would be very thankful for your help, or if anybody else in the forum can jump in to clarify my doubts.

Thanks a lot!

To be clear, I want to get an impulse response function of consumption to government spending, when agents live in an economy in which they know that the interest rate will be at the lower bound for 10 periods, but will then follow a Taylor rule that is also known.

Should I use the forecast function? I’ve tried, but I get results that make no sense, extremely low and not changing values for endogenous variables.


Any help would be greatly appreciated!

we’ve been trying to do a similar exercise. and we used stoch_simul along with the forecast command.

and at this point we’re not sure how the forecast command works. in a fully deterministic model, we were able to get the interest rate=0, but not in a stochastic model. so maybe the forecast command is not doing what we think it is doing.

were you able to figure out a way?