Tracing Error Message

[bigbigben]

My model runs well for the computational part, but it doesn't work for estimation. Actually, I just tried one series and estimating one parameter.
My specification about estimation is as follows:

varobs gy_obs;
estimated_params;
rhor,normal_pdf,0.6,0.1;
end;
estimation(datafile=TSJH_data,nobs=205,loglinear,mh_replic=0,mh_nblocks=5,mh_jscale=.;


The following is what I got.

??? Error using ==> dynare_estimation
There are complex values in the data. Probably a wrong transformation

Error in ==> TSJH_est_Oct9 at 216
dynare_estimation(var_list_);

Could anyone tell me what default transformation is done in Dynare, since the original data are just growth rate of GDP and the scale is 0.01.

[MichelJuillard]

When you ask for loglinear approximation, Dynare takes the log of the variables. So you shouldn't take transform the data before hand or you shouldn't use a loglinear transformation if some of your data take a zero or a negative value.

Best

Michel