In my model, I have the following equation in which A is the technology:
A = A (-1) ^ rhoA * exp (epsA)
In estimating, I wrote the following:
stderr epsA, inv_gamma_pdf, 0.01, inf;
But after estimating, the following answer was obtained for epsA:
posterior_mode: value of epsA=4.2931
posterior_std_at_mode: value of epsA=5.1575
Is my model wrong? Or can such a result be achieved and not a problem?
That looks a bit strange. Did you multiply your data by 100? Does your prior reflect this?
No, I haven’t multiplied my data 100 times, and my predecessor doesn’t reflect that. Also graph of MCMC univariate convergence diagnostic (Brooks and Gelman, 1998) for epsA is as follows:
As a result, the Multivariate convergence diagnostic diagram is as follows:
Without knowing the model it is impossible to tell. It may simply be that for your country/sample the data is very volatile. Or there is still a mistake somewhere.
If only this variable has a problem, are the results reliable?
Again, you are the model builder. You should have an idea about whether the results make sense.