Help with declaring model local variables

[Nice]

Hello All,

1. I am estimating a DSGE model and wanted to verify whether I have correctly specified model local variables. I used " A Guide to Specifying Observation Equations" to guide me but am still unclear if I have done so correctly.

2. The estimation runs however I am not sure if the estimation results are correct. I am replicating the model and when I compare the estimation results in Dynare with the ones found in the paper some parameter values are different. For example the habit formation parameter estimate is 0.14 where as in the paper it is 0.5, does this suggest there is an error in the estimation?

Many Many Thanks!

[jpfeifer]

The model-local variables seem to be OK, but you cannot estimate parameters that are in the shocks-block. If you define

Code: Select all

shocks;
var e_u_b_h  = SIGMA_B_H  ^2;
var e_u_l_h  = SIGMA_L_H  ^2;
var e_u_i_h  = SIGMA_I_H  ^2;
var e_u_a_h  = SIGMA_A_H  ^2;
var e_u_g_h  = SIGMA_G_H  ^2;
var e_u_z_h  = SIGMA_Z_H  ^2;
var e_u_tk_h = SIGMA_TK_H ^2;
var e_u_tl_h = SIGMA_TL_H ^2;
var e_u_tc_h = SIGMA_TC_H ^2;
end;


this only works for calibration. You cannot estimate SIGMA_B_H. Rather, you must estimate

Code: Select all

stderr e_u_b_h,  inv_gamma_pdf, 0.01, inf;


[Nice]

Thank you so much jpfeifer!! Your help is much appreciated.

Kind Regards,

Nice