Estimating the correlation between shocks

Dear all,
I got some troubles using Dynare 4 by estimating the correlation between shocks. So in the “shocks” section I just assumed that the correlation coefficient between shocks is equal to 1:

var e_a; stderr 0.5;
var e_a, e_x = 0.50.5;
var e_ph; stderr 0.5;
var e_ph, e_x = 0.5
var e_t; stderr 0.5;
var e_t, e_x = 0.50.5;
var e_th; stderr 0.5;
var e_th, e_x = 0.5
var e_x; stderr 0.5;

In the “estimated_params” section I specified correlation coefficients to be estimated using MLE as follows:
stderr e_a,0.5,;
corr e_a, e_x,0.123,-1,1;
stderr e_ph,0.5,;
corr e_ph, e_x,0.152,-1,1;
stderr e_t,0.5,;
corr e_t, e_x,-0.845,-1,1;
stderr e_th,0.5,;
corr e_th, e_x,-0.427,-1,1;
stderr e_x,0.5,;

After running the .mod-file I got estimation results for standard deviation and correlation coefficients of specified shocks
AND the following error message:

??? Undefined command/function ‘Sigma_e_’.

Error in ==> set_all_parameters at 42
Sigma_e(k1,k2) = xparam1(i+offset)*sqrt(Sigma_e_(k1,k1)*Sigma_e_(k2,k2));

Error in ==> DsgeSmoother at 35

Error in ==> dynare_estimation at 887
[atT,innov,measurement_error,filtered_state_vector,ys,trend_coeff] = DsgeSmoother(xparam1,gend,data);

Error in ==> test12_2 at 284

Error in ==> dynare at 79
evalin(‘base’,fname) ;

Until now I have not found out, why this error message shows up.
I would be glad if somebody knows about this error and could explain me how to deal with that.
Thank you!

We just corrected bugs relative to estimation of models with measurement error. Please, grab the last snapshot.
Thanks for reporting the problem