Dear all,

I have four variables, one of which is the price of crude oil and the other three variable are three different freight rates.
I want to analyze the relationship between these four variables, so I try to use VAR or VECM model.

I divide the sample into
1. A sample of all time (from 2017/1/1 ~ 2022/12/31)
2. Before the outbreak of pandemic (from 2017/1/1 ~ 2019/12/31)
3. After the outbreak of pandemic (from 2020/1/1 ~ 2022/12/31)

After the Johansen co-integration test, it was found that they all have co-integration relationship, so I used the VECM model.

However, when I use the command

Code:
vecnorm, jbera
the three samples all reject the null hypothesis, which means that the samples do not conform to normality.

And the command
Code:
vecstable
shows The VECM specification imposes a (or 2) unit moduli.

Code:
veclmar
also shows rejection no autocorrelation at lag order


These problems have troubled me for a long time. How can I fix these issues?


Thanks in advance!