Hello , i'm about to study volatility spillovers between Index (A and C) , (B and C) , i can chose whether a bi- variate model for the study between (A and C) , (B and C) or a tri variate model GARCH and study the volatility spillovers between the tree index at once , i would like to learn which is the best ? and what is the difference between the two methodologies ? and thanks Again
.Dear Boubaker Touijrat
The best modle is the model has lowest values of the following criterias: Akaike's information criterion (AIC), the Bayesian information criterion (BIC) and the Hannan-Quinn cri- terion (HQ).
I am assuming that your models are predicting daily volatilities. In addition to what Abbas says above (which I agree with), you can also use average squared prediction error, if you have tick data for intra-day index values available and can hence calculate realized volatilities (as daily standard deviations) from every 15 min readings, say.
- Badges
- Science topic
- Similar topics
- Psychology
- Mental Processes
- Learning