Panel VAR Diagnostics / Assumptions.

[John Riveros]

Greetings Statalist Users and team.

I'm working on a panel data VAR model using pvar package. Where I'm assuming some macroeconomic variables are endogenous. I'm using first differenced data in order to have stationarity (ips and llc tests confirmed that variables are stationary at first differences). I did not have trouble estimating the model with the following syntaxis.

pvarsoc d_ln_y d_ln_s d_ln_amp d_ln_m         //lag 3
pvar d_ln_y d_ln_s d_ln_amp d_ln_m, lag(3)
pvargranger
pvarstable
pvarirf
pvarfevd

The model took as instruments the lags (1/3) of my variables.

Now, I'm wondering how can I test the assumptions of no cross-sectional/contemporaneous correlation, VIF, no serial correlation, heteroskedasticity and such?.

Thank you for your time.

[John Riveros]

No information regarding this topic?

[Jenner Kerr]

?

[Nick Cox]

https://www.statalist.org/forums/help#adviceextras #1 explains Statalist etiquette on bumping.

This is one of many topics on which there are more questions than answers. The explanation is probably no deeper than

1. The program authors or technique experts just don't hang out here. Whenever I search this topic I end up with one name above all and that person is not active here.Their choice, absolutely.

2. The people who do hang out here and answer many questions have little or no experience of the technique or the command in question

You could try emailing direct but I can make no promises on anybody's behalf. Many people take the line that being an expert on something doesn't create the time or inclination to support everyone or even anyone who has a question on the topic. Indeed, this is my line on most emails or PMs I receive directly.

[John Riveros]

Thank you a lot Nick, I tried contacting the author, however, it does not have a response so far. Yet, I kept digging and found some useful stuff which had to be done manually after post-estimation. Happy new Year, and merry christmas!

[Nandini Suresh]

Hi John, what is the procedure to be done manually after post-estimation?