I'm running a VEC model and get Durbin-Watson statitics where d < dL,α, wich means there is statistical evidence that the error terms are positively autocorrelated. However, a Ljung-Box Q test with 5 lags (based on rule of thumb out of sample size) can't reject H0 that data are independently distributed. What test should I trust? Correlogram of residuals from the second equatio looks ok (attached file). What should I do? Add more lags?
s. Béatrice Marianne Ewalds-Kvist
Check if this helps you:
http://stats.stackexchange.com/questions/148004/testing-for-autocorrelation-ljung-box-versus-breusch-godfrey
1 votes
0 thanks
- Badges
- Science topic
- Similar topics
- Medicine
- Physiology
- Biosignals
- Biomedical Signal Processing
- Statistical Analysis
More Lars Ahnland's questions See All
Similar questions and discussions