I'm working on time series analysis in R, specifically using the VAR model to understand the relationship between different variables like "GPRD_THREAT" and the price of wheat ("WHEAT"). I've used the vars package for this purpose.
After fitting the model with different lags and running the serial correlation test using serial.test(), I'm consistently getting p-values less than 2.2e-16 and zero degrees of freedom, which doesn't make sense. How can I correctly interpret these test results, or could there be an issue with my implementation?
tried running the serial test on VAR models with different lags. I was expecting to find a model that passes the serial test, i.e., has no serial correlation in the residuals. This would indicate that my model is well-specified. However, regardless of the lags I used, the serial tests are failing. I'm unsure whether this is due to a problem in my code, the data, or my understanding of the test.