Dynare forums

There still seems to be a mistake in the osr command in the latest (stable) version of Dynare. (I am using 4.4.2 on a Mac at home, and version 4.4.2 on Windows at work.)

The problem is that when a covariance is added in the welfare or loss function (as in the mod example below), the preprocessing of the mod file into the Mat file yields non-unique elements in obj_var_. This vector is passed to the ors function (the vector is renamed i_var) and then gets passed to the osr1 function which calls osr_obj (passing a reordered i_var vector to ors_obj), which then computes the welfare function by computing variances and covariances using the following two lines of code:

vx = get_variance_of_endogenous_variables(dr,i_var);
loss = weights(:)'*vx(:);

In the example below vx should be 3 by 3, but it ends up being a 4 by 4 matrix, conformable with the 4 by 4 weights matrix.
full(weights)
ans =
1.0000 0 0 1.0000
0.5000 1.0000 1.0000 0.5000
0.5000 1.0000 1.0000 0.5000
1.0000 0 0 1.0000

The weights matrix gets conformably messed up in the call to ors_obj (line 61 of osr1):
[loss,vx,info,exit_flag]=osr_obj(t0,i_params,inv_order_var(i_var),weights(i_var,i_var));

The bug can, I believe, be fixed by adding the following at the beginning of the osr function
i_var=unique(i_var);

Though ideally the dynare prepocessing would apply the unique command to obj_var_

Regards,
Christie


% Example Mod file taken from http://www.dynare.org/manual/index_29.html

var y inflation r;
varexo y_ inf_;

parameters delta sigma alpha kappa gammarr gammax0 gammac0 gamma_y_ gamma_inf_;

delta = 0.44;
kappa = 0.18;
alpha = 0.48;
sigma = -0.06;

gammarr = 0;
gammax0 = 0.2;
gammac0 = 1.5;
gamma_y_ = 8;
gamma_inf_ = 3;

model(linear);
y = delta * y(-1) + (1-delta)*y(+1)+sigma *(r - inflation(+1)) + y_;
inflation = alpha * inflation(-1) + (1-alpha) * inflation(+1) + kappa*y + inf_;
r = gammax0*y(-1)+gammac0*inflation(-1)+gamma_y_*y_+gamma_inf_*inf_;
end;

shocks;
var y_; stderr 0.63;
var inf_; stderr 0.4;
end;

optim_weights;
inflation 1;
y 1;
y, inflation 0.5;
end;

osr_params gammax0 gammac0 gamma_y_ gamma_inf_;
osr y;

Correction: vx should be 2 by 2, not 3 by 3 in the example I provided.
Christie

We are going to look into this. But this seems to be no problem, because the objective function is just multiplied by 4, meaning that the minimum stays the same.