Dear all,
I am trying to develop a model with heterogeneous monopolistic firms and search & matching frictions, however I cannot achieve a stable equilibrium. By reading previous posts on this topic I understand that there should be some problem in the timing, however I still cannot figure it out. May someone help me?
Thanks in advance,
best
Marco
search_monopol.mod (6.1 KB)
SS.mat (1.2 KB)
I cannot spot anything immediately suspicious, but I also don’t know the model. Have you tried whether different parameterizations solve the problem? Otherwise, you can only simplify the model further.
Dear Professor Pfeifer,
thank you for the answer. I discovered some mistakes I made in setting initial values for the steady state. Now the model works, however the impulse response functions for a positive TFP shock display some disturbance in the first periods of simulation; again, I wonder whether it might be due to wrong timing that I cannot spot. May I ask you your opinion about it?
Thank you for your time
Marco
search_monopol.mod (6.5 KB)
SS.mat (1.2 KB)
Such oscillations are indeed usually a sign of a timing error (giving rise to a complex eigenvalue). However, if there are multiple steady states, this can also happen if you pick the wrong one.
Dear all,
I simplified the model and I corrected some mistakes, falling again into the problem of unstable equilibrium. By running model diagnostics Dynare found collinearity between the Job Creation Condition and the wage schedule; by removing search frictions a stable equilibrium is found and the collinearity issue disappears, whereas by removing only heterogeneity Dynare is still able to find a stable equilibrium but collinearity remains. May I ask some help/suggestions?
As always, thank you in advance.
Best
Marco
search_monopol.mod (9.8 KB)
SS.mat (1.7 KB)
Your mat-file does not contain the N_ss required by the mod-file.
Dear Professor Pfeifer,
I apologize, the file was an older version. Anyway I was able to solve the problem: the dynamics for aggregate vacancies depending on entrants and incumbents were misspecified, causing both errors. Now the model seems to work.
Thank you again for your time and insights,
best
Marco
