Simulation error dynare 5.0

Dynare++
1

Hi,

I’ve been trying to replicate Table 3 of Croce (2014) using the code provide by the author, but when I tried to simulate the series with the following code:

per = 840; %number of simulated months
n   = 100; %number of simulations

load Approx_10.mat;

shocks = zeros(2,per)./0; % put NaNs everywhere
r      = zeros(28,per,n);

for i=1:n
    r(:,:,i)  = dynare_simul('Approx_10.mat',shocks);
end

I received the error:

Error using dynare_simul_
dynare_simul_ must have at exactly 12 input parameters and 1 output argument.

Error in dynare_simul (line 171)
[err,r]=dynare_simul_(order-1,nstat,npred,nboth,nforw,…

Error in croce2014table3 (line 20)
** r(:,:,i) = dynare_simul(‘Approx_10.mat’,shocks);**

I “solve” the problem editing dynare_simul.m in the following way:

[err,r]=dynare_simul_(order-1,nstat,npred,nboth,nforw,...
    nexog,ystart,shocks,vcov_exo,seed,ss,dr);

if err
    error('Simulation failed')

for

% call dynare_simul_
[r]=dynare_simul_(order-1,nstat,npred,nboth,nforw,...
    nexog,ystart,shocks,vcov_exo,seed,ss,dr);

%if err
   % error('Simulation failed')

My questions: is this affecting my results? Are there another way to avoid the problem? Also I’m using dynare 5.0.

Approx_10.mod (4.5 KB)

2

I would not use Dynare++, but go directly for Dynare 5.0:
Approx_10.mod (4.7 KB)

3

But the author uses explicitly Dynare++, this change won’t affect the results?

4

If you want to be sure to obtain the same results as the original authors, you need to use the same Dynare++ version. If you want to have correct results in the sense of the lower number of bugs potentially affecting the results, use Dynare 5.0.