Please help on running a code

[irinka]

Hi again,

I am trying to make an estimation using the MH algorithm and I was wondering if you can help as I have encountered some errors.
Based on the model of De Graeve (2008), I used data for the US economy.

Can you please help me to fix these type of errors? I have used the same calibration as in his working paper.
Here I attached the .mod file and the data used. I am very curious to make it work!

Thank you!

[irinka]

Please, can you help me? I really want to understand the mechanism and I cannot figure out how can I solve those errors.

I would really appreaciate your help!

[jpfeifer]

1. You are using the loglinear option. This will take the log of the data unless you use the logdata option (only available in the unstable version).
2. You have 7 observables, but only 6 shocks with non-zero standard deviation. You must have at least 7, i.e. as many shocks as observables to prevent stochastic singularity.
3. You priors are incompatible with the model. E.g.

Code: Select all

KbarNbar, uniform_pdf, 0, 5; 

i.e. the capital-labor ratio has mean 0 and can be negative. That is wrong. Make sure the model runs with calibrated values (which it currently does not due to Blanchard-Kahn violations) and then use

Code: Select all

estimated_params_init(use_calibration);
end;