Occbin in Dynare 5.0

Hi, I am trying to simulate a DSGE Model with ZLB using Dynare 5.0 and Occbin block. My question is how to do stochastic simulation and get for example the unconditional mean of production function (y) or Welfare function (U);

var c r k l w pi r_k y mc i z U;

varexo eps_z;

parameters varphi beta delta alpha phi epsilon rho_r phi_pi phi_y rho_z sigma;

varphi = 1.5;

beta = 0.99;

delta = 0.025;

alpha = 0.33;

phi = 106;

epsilon = 6;

rho_r = 0.6;

phi_pi = 1.5;

phi_y = 0.125;

rho_z = 0.95;

sigma = 0.01;

model;

#z_ss = 0;

#pi_ss = 0;

#r_ss = 1/beta;

#l_ss = 1/3;

#r_k_ss = 1/beta - (1 - delta);

#mc_ss = (epsilon-1)/epsilon;

#w_ss = (mc_ss*(alpha/r_k_ss)^alpha)^(1/(1-alpha))*(1-alpha);

#k_ss = w_ss/r_k_ss*alpha/(1-alpha)*l_ss;

#y_ss = exp(z_ss)*k_ss^alpha*l_ss^(1-alpha);

#i_ss = delta*k_ss;

#c_ss = y_ss - i_ss;

#chi = w_ss/c_ss*l_ss^(-varphi);

//Welfare

U = ln(c) - l^(1+varphi)/(1+varphi) + beta*U(+1);

//Euler equation

beta*c/c(+1)*r/(1+pi(+1)) = 1;

//Labor supply

chi*l^varphi = w/c;

//Capital supply

beta*c/c(+1)*(r_k(+1) + 1 - delta) = 1;

//Capital law motion

k = (1-delta)*k(-1) + i;

//Production function

y = exp(z)*k(-1)^alpha*l^(1-alpha);

//Marginal cost

mc = (w/(1-alpha))^(1-alpha)*(r_k/alpha)^alpha;

//Labor capital ratio

w/r_k = (1-alpha)/alpha*k(-1)/l;

//Phillips curve

phi*pi*(1+pi) = 1 +epsilon*(mc-1) + beta*(c/c(+1)*phi*pi(+1)*(1+pi(+1))*y(+1)/y);

//Resource constraint

(1-phi/2*pi^2)*y = c + i;

[name='Taylor rule',relax = 'ZLB']

r = r_ss + phi_pi*pi + phi_y*(y-y_ss);

[name='Taylor rule',bind = 'ZLB']

r = 1;

//TFP AR(1)

z = rho_z*z(-1) +eps_z;

end;

occbin_constraints;

name 'ZLB'; bind r<1;

end;

steady_state_model;

z = 0;

pi = 0;

r = 1/beta;

l = 1/3;

r_k = 1/beta - (1 - delta);

mc = (epsilon-1)/epsilon;

w = (mc*(alpha/r_k)^alpha)^(1/(1-alpha))*(1-alpha);

k = w/r_k*alpha/(1-alpha)*l;

y = exp(z)*k^alpha*l^(1-alpha);

i = delta*k;

c = y - i;

U = (ln(c) - l^(1+varphi)/(1+varphi))/(1-beta);

end;

//----------------------------------------

// Shocks

//----------------------------------------

shocks;

var eps_z; stderr 0.1;

end;

steady;

occbin_setup(simul_maxit = 100, simul_periodic_solution, simul_debug);

Stochastic simulations are triggered by a call to occbin_solver;

Thank You for Your response, but using occbin_solver without shocks(surprise); returns me the endogenous variables on its mean value. And using shock(surprise) is like doing deterministic simulations in the way that I have to specify the size of shock for each simulation period. I want to do a simulation where I give the Dynare the mean std of the shock and its simulates the model for specified periods.

Is this what you are after?
model.mod (2.3 KB)

Yes, thank You

Also, one more question, why at some point (in Your .mod file it is 31 period) the constraint doesn’t work and interest rate is less than 1 (ZLB constraint)?

model.mod (2.4 KB)

The problem was that the implementation of the ZLB was wrong.
Therefore, you had the issue of a periodic solution. Now it works without the option simul_periodic_solution.
Most of the time a periodic solution is an indication of an implementation bug like the one at hand.

Thanks

7 posts were split to a new topic: Problem with ZLB implementation

2 posts were split to a new topic: QE model with Occbin

7 posts were split to a new topic: ZLB in two-country model