Code running into problem

Estimation ML/Bayesian estimation
21

See

22

Thank you for your reply Prof. Pfeifer. My code could run for about half an hour, then an error showed up stating the Error using chol. When I changed the prior variance to a small number, the code could not run at all. The two problematic parameters follow Beta distribution. One is the monetary policy persistence, and the other is the governemnt expenditure persistence. I do not know why the code cannot run at all when I changed the variance to a smaller number.

23

Without the codes, it is impossible to tell. What is the exact error message?

25

There is a problem with your model:

MODEL_DIAGNOSTICS: The following endogenous variables aren’t present at the current period in the model:
dbar

26

Thank you Prof. Peifer, for your reply. I have two more questions. First, does all endogenous variables have to present at the current period in the model? Second, my calibrated model still works. I am confused why it is still works if there is a problem with the model. I am looking forward to your reply. Thank you again. Alex

27
  1. I am unaware of sensible models where variables do not appear in the current period. After all, how are you supposed to determine the past, predetermined values if there is no contemporaneous decision about them?
  2. You may be able to run the model. But that does not mean it makes sense.
29

Then you should try to simplify your model. It’s hard to debug a model of that size.

30

Thank you, Prof. Pfeifer, for your advice. I have one question about the dynare forum. Sometimes when you reply to my message, I receive an email notification. But sometimes I do not receive a notification. What should I do so that I can receive a notification when others reply to my questions.

31

At the bottom of the page, you can set the “Tracking” of threads.

32

Thank you, Prof. Pfeifer, for your reply. Alex

33

Even if there is a problem within the model, it is still possible to get the calibrated results although the results may do not make sence at all. I am curious if it is possible to get the estimation results if there is a problem witin the model. Thank you Prof. Pfeifer.

34

Usually, if you cannot get the model to run properly by choosing parameters, trying to estimate it will fail.

35

Thank you Prof. Pfeiper, for answering my question.

36

Dear Prof. Pfeifer,

I have the same warning "Warning: gmhmaxlik: Unknown option (MaxIter)! ". However, I can get the results. Do I need to deal with this warning?

Thank you,
Alex

37

Without the codes and the full message it is impossible to tell. But the message seems safe to ignore.

38

Thank you Prof. Pfeifer for answering my question.