Dynare forums

Dear all,
I am just doing my first steps in Dynare. I have a system comprising of a new keynesian phillips curve, an aggregate demand and a money demand. I estimated this system in some econometric software and I want to solve it in Dynare and shock money permanently. Now any such model will predict that the price level in the end of the period will have an equiproportional increase with the money shock (i.e. if a 10% increase in money is introduced the price level jumps up by 10%). I get all other results-responses fine but not this one. Below I write my module. Any thoughts? Warmest thanks in advance.


var p, y, tbill;
varexo yp, m1, unkpc, uad, umd;

parameters c1, c2, c3, c4, c5, c6, c7, c8, c9;

c1 = 0.4252097;
c2 = 0.5506059;
c3 = 0.0035550;
c4 = 0.0024613;
c5 = -0.3728166;
c6 = -0.0292797;
c7 = 0.0394045;
c8 = 0.5395523;
c9 = 0.1652910;

model;
p = p(-4)+c1*(p(+1)-p(-3))+c2*(p(-1)-p(-5))+c3*(y-yp)+unkpc;
y = c4+y(+1) + c5*tbill-c5*(0.435748*(p(+1)-p(-3))+0.564252*(p(-1)-p(-5)))+uad;
m1 = p+m1(-1)+p(-1)+c6*(m1(-1)-p(-1)+1.996070-0.490928*y(-1))+c7*tbill(-1)+c8*(m1(-1)-p(-1)-m1(-2)+p(-2))+c9*(m1(-2)-p(-2)-m1(-3)+p(-3))+umd;
end;

initval;
p = 4.058371902;
y = 8.731896101;
tbill = 0.09516666667;
yp = 8.656415858;
m1 = 6.245129422;
unkpc = -0.00010429110196597;
uad = -0.00125188907392726;
umd = -0.0119632003673317;
end;
steady;

endval;
p = 4.058371902;
y = 8.731896101;
tbill = 0.09516666667;
yp = 8.656415858;
m1 = 6.255129422;
unkpc = -0.00010429110196597;
uad = -0.00125188907392726;
umd = -0.0119632003673317;
end;
steady;

shocks;
var m1 = 0.01;
end;

stoch_simul(periods=2000);

Most probably your model is still wrong. Your model clearly should have a unit root because a shock to money should permanently move the price level and hence the steady state of the model. This is the effect you are looking for. However, if you put a check; command after steady, you will see that there is no eigenvalue of 1. Hence, there are no permanent effects in the model you entered into Dynare.

Hi jpfeifer,
Warmest thanks for the turbo reply. Any specific thoughts/ suggestions in terms of altering the model and address this issue would be more than appreciated. That is of course if I do not become a burden on your time.

Best
M

The model has become:
var p, y, tbill;
varexo yp, m1, rnkpc, rad, rmd;

parameters c1, c2, c3, c4, c5, c6, c7;

c1 = 0.809334;
c2 = 0.1091944;
c3 = 0.004095;
c4 = 0.0014805;
c5 = 0.0887723;
c6 = -0.0569114;
c7 = 0.1321844;

model;
p = p(-1)+c1*(p(+1)-p)+c2*(p(-1)-p(-2))+c3*(y-yp)+rnkpc;
y = c4+y(+1)-c5*tbill+c5*(0.435748*(p(+1)-p)+0.564252*(p(-1)-p(-2)))+rad;
m1 = p+m1(-1)+p(-1)+c6*(m1(-1)-p(-1)+1.996070-0.490928*y(-1))+c7*tbill(-1)+rmd;
end;

initval;
p = 4.0220604;
y = 8.676991;
tbill = 0.1304333;
yp = 8.62639;
m1 = 6.1139;
rnkpc = 0.0038666;
rad = 0.0132505;
rmd = -0.0299652;
end;
steady;

endval;
p = 4.42426644;
y = 8.676991;
tbill = 0.1304333;
yp = 8.62639;
m1 = 6.2139;
rnkpc = 0.0038666;
rad = 0.0132505;
rmd = -0.0299652;
end;
steady;

shocks;
var m1=0.1;
end;

stoch_simul(periods=200, irf=100);