Dear forum:
I know this is not a new topic and I have searched some posts in the forum. I’m still a little confused about how to do the simulation for the unrealized news shock.
For example,
/A benchmark NK Model/
var y c m r i pi n w k mc rk lambda a v Q I;
varexo es_a es_v_surprise es_v_news;
parameters beta sigma phi nu alpha delta RSS KSS CSS phi_pi phi_y phi_rho;
parameters thet psi rho_a sigma_a rho_v sigma_v;
beta = 0.99;
sigma = 1;
phi = 1;
nu = 1;
alpha = 1/3;
delta = 0.025;
RSS = 1/beta;
KSS = alpha/(RSS-1+delta);
CSS = 1 - delta*KSS;
phi_pi = 1.5;
phi_y = 0.5;
phi_rho = 0.8;
thet = 0.8;psi = 4;
rho_a = 0.8;
sigma_a = 0.1;
rho_v = 0.8;
sigma_v = 0.1;model(linear);
%Euler Equation
-sigmac = lambda;
%Labor supply
phin = lambda + w;
lambda = lambda(+1) + r;
%Real money balance
-num - lambda = ibeta/(1-beta);
%Fisher Equation
i = r + pi(+1);
Q = deltapsi(I - k(-1));
k = (1-delta)k(-1) + deltaI;
RSS*(Q + r) = (RSS-1+delta)rk(+1) + (1-delta)Q(+1) + psideltadelta*(I(+1)-k);
%Production function
y = alphak(-1) + (1-alpha)n + a;
%factor return, from FOC on profit maximization
w + n = k(-1) + rk;
%Real marginal cost
mc = (1-alpha)w + alphark - a;
%Phillips Curve
pi = betapi(+1) + (1-thet)(1-thetbeta)mc/thet;
%Market clear condtition
y = CSSc + KSS k - (1-delta)KSSk(-1);
%Taylor Rule
i = phi_rhoi(-1) + (1-phi_rho)(phi_pipi(+1) + phi_yy(+1)) + v;
%Technology shock
a = rho_aa(-1) + es_a;
%Monetary policy shock
v = rho_vv(-1) + es_v_surprise+es_v_news(-4);
end;
steady;
check;
shocks;
var es_v_news; stderr 0.1;
var es_v_surprise = sigma_a^2;
end;
stoch_simul(irf=20,print)y c m r i pi n w k;
How can I make the surprise shock es_v_surprise to offset the anticipated shock es_v_news at time t=0 so es_v_news wont be realized? The two shocks are already the same size. Can I just do it like:
v = rho_v*v(-1) - es_v_surprise+es_v_news(-4);
I guess case wont be so easy.
I have read a code in
https://archives.dynare.org/phpBB3/viewtopic.php%3Ff=1&t=4802.html
but in the above linearized model, results in oo_.dr.ys will be 0 since steady state vars are 0.
So how can I do the simulation for an unrealized news shock in a log-linearized model?
Can anybody give me some hints?
Thank you sincerely!