Aniello Ferraro

Dears, I have a problem in step 5. Once I have set the various settings I get this error message: How could I solve? 5) IMPULSE RESPONSE ANALYSIS 5.1) Settings, shock selection and (optional) region definition for regional GIRFs and GFEVDs 5.2) Computing the covariance matrix Warning: The covariance matrix is not positive definite. Pause and go to gvar1.xls: Ensure that perform shrinkage on the correlation matrix for generating the bootstrap data is set to 1 (if it is not already), then press enter....

I solved the problem. In the contiguity matrix that I loaded there were some regions without neighbors (it was a queen matrix, so for the islands there was the value 0 for all the countries) eliminating those units the problem was solved. I hope it will be of help to someone

Thank you very much for the reply. I double-checked the guide and the data entered in the excel gvar.xls file, but there must be something I'm missing. In my file I have 20 countries (not grouped by regions), 5 country-specific variables, 1 global variables, and I manually entered my weight matrix. The problem arises in phase 3.1) of the Model specification, whether the dominant unit model is enabled or not. If I do not enable the dominant unit model, as required by the program, I insert the global...

Dears, in phase 3.5 I always get this error message both if I select the aic and sbc criterion. How can I fix? Thanks to those who will help me 3.5) Determining the lag orders of each country model Error using rank Input matrix must not contain NaN or Inf values. Error in AIC_SBC (line 20) s = rank(X);