stepan
/
dynare
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Computation of the loss function in the presence of unit root variables was incorrect 

Some elements were simply left out. This commit now mimics the computations as in th_autocovariances.m and makes sure the loss function coincides with the one that is obtained by reading out the covariance from oo_.var
time-shift
Johannes Pfeifer 2014-11-03 18:05:15 +01:00
parent 669016167e
commit cb5d3df6a0
1 changed files with 2 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

7
matlab/get_variance_of_endogenous_variables.m
View File

@ -48,16 +48,13 @@ n = length(i_var);
[vx,u] = lyapunov_symm(A,B*Sigma_e*B',options_.qz_criterium,options_.lyapunov_complex_threshold, [], [], options_.debug);
if size(u,2) > 0
i_stat_0 = find(any(abs(A*u) < options_.Schur_vec_tol,2));
i_stat = find(any(abs(ghx*u) < options_.Schur_vec_tol,2));
i_stat = find(any(abs(ghx*u) < options_.Schur_vec_tol,2)); %only set those variances of objective function for which variance is finite
ghx = ghx(i_stat,:);
ghu = ghu(i_stat,:);
else
i_stat_0 = 1:size(ghx,2);
i_stat = (1:n)';
end
vx1 = Inf*ones(n,n);
vx1(i_stat,i_stat) = ghx(:,i_stat_0)*vx(i_stat_0,i_stat_0)*ghx(:,i_stat_0)'+ghu*Sigma_e*ghu';
vx1(i_stat,i_stat) = ghx*vx*ghx'+ghu*Sigma_e*ghu';