Shocks for multiple periods

Stochastic simulations
1

I’m forecasting macro variables for several countries when we impose shocks on variables such as the unemployment rate, federal funds rate, etc. Currently, I am able to impose 1 period shocks. Example, if the equation for federal funds rate (defined by a Taylor rule) is measured in % terms as follows:

fed_funds_rate = linear combination of variables + error_ffr

Then, to impose an 80 % upward shock on the rate, I code:
var error_ffr; stderr 0.8

But this is only for one period. How can I extend it to last for more than 1 periods?

I tried

var error_ffr;
periods 1:4;
values 10;

But I don’t think this is the right way. I would also like to save these impulse responses evaluated for multiple periods for other usage in MATLAB.

2

Are we talking anticipated or surprise shocks in each period? See also

3

Thank you. That helps

4

By the way, if I write the standard error for two different variables, example:

var error_ffr; stderr 0.8;
var error_inflation_rate 0.7;

resid;
steady(maxit = 700, solve_algo = 1);
stoch_simul(order=1, irf = 40, graph) names_of_variables

Will the impulse response graphs measure the impact due to both the aforementioned shocks on inflation rate and the federal funds?

5

No, you will get two sets of IRFs, one for each shock.

6

Do you have an example of Dynare code to impose multiple “surprise” shocks for multiple periods?

So far, I have imposed multiple shocks on the inflation rate and federal funds rate for one period only by creating a third exogenous variable; applying a scaling factor to it, and adding it to the equations that define the inflation rate and the federal funds rate. Then, I specified the variance of this third exogenous variable. The output is the impulse response for 2 simultaneous shocks imposed in one period.

I also explored the following which is mentioned in page 32 of the Dynare manual
var residual_inflation
periods 1:8
values 3

var residual_ffr
periods 1:8
values 5

resid;
steady;
stoch_simul(order = 1, irf = 30, graph)ffr inflation_rate output_gap

However, this doesn’t generate the IRFs. Why does it not produce the IRFs? I know you mentioned that simult_function can generate IRFs for anticipated shocks, but I wonder if we can modify the above snippet of code I just mentioned so that it also produces the desired IRFs ?

7

Your syntax above is for perfect foresight simulations, but you are doing stochastic simulations. You need to use the simult_-function as in the linked example.