Identification Problem and Dynare Error Message DSGE model

[LSHui]

Hi, all
I am trying to use a DSGE model based on Adolfson 2005 to do forecast.
When I run identification command, there are 25 parameters can not be identified by model, after go through Dynare forums and found solution from this topic: viewtopic.php?f=1&t=4982&p=12685&hilit=quasi+bug#p12685, 19 of them can be identified after following jpfeifer's advice. But I cannot figure out the identification problem in the rest 6 of them, which are :

xi_w ----- Probability with which a worker cannot reoptimize his wage
kappa_w ------Indexation parameter for wages
lambda_m_c_ss ---- Steady state time-varying markup in the imported consumption goods sector
lambda_m_i_ss ---- Steady state time-varying markup in the imported investment goods sector
tau_w_ss ------- Steady-state payroll tax
tau_y_ss ------- Steady-state labour income tax

could you please give me some advice about them? What will cause if we ignore identification problem and do estimation anyway?
Much appreciated!!

Helen

[jpfeifer]

The problem is that those parameters enter composite parameters that are once set to a calibrated value and never updated. For example, you try to estimate xi_w. xi_w enters in e.g.

Code: Select all

b_w=(lambda_w*sigma_L-1+lambda_w)/((1-beta*xi_w)*(1-xi_w));

But you are only providing the value for b_w to Dynare while not telling Dynare that b_w depends on xi_w. See the document and Remark 4 in https://sites.google.com/site/pfeiferecon/Pfeifer_2013_Observation_Equations.pdf?attredirects=0 for a clarifying example.

[LSHui]

Thank you very much, jpfeifer.

Your reply is so quick,very clear and helpful, especially the linked document. I will follow your advice to figure out the problem.
By the way, could you please sharing more documents about dynare just like the one you linked in previous post, because it likes a textbook about dynare and DSGE model, practically and systematically.

Thank you for your time!


Helen